{"syncedAt":"2026-07-07T02:10:18.945Z","providers":{"requesty":{"id":"requesty","name":"Requesty","doc":"https://requesty.ai/solution/llm-routing/models","npm":"@ai-sdk/openai-compatible","env":["REQUESTY_API_KEY"],"nm":38},"qiniu-ai":{"id":"qiniu-ai","name":"Qiniu","doc":"https://developer.qiniu.com/aitokenapi","npm":"@ai-sdk/openai-compatible","env":["QINIU_API_KEY"],"nm":91},"alibaba-cn":{"id":"alibaba-cn","name":"Alibaba (China)","doc":"https://www.alibabacloud.com/help/en/model-studio/models","npm":"@ai-sdk/openai-compatible","env":["DASHSCOPE_API_KEY"],"nm":84},"regolo-ai":{"id":"regolo-ai","name":"Regolo AI","doc":"https://docs.regolo.ai/","npm":"@ai-sdk/openai-compatible","env":["REGOLO_API_KEY"],"nm":13},"stackit":{"id":"stackit","name":"STACKIT","doc":"https://docs.stackit.cloud/products/data-and-ai/ai-model-serving/basics/available-shared-models","npm":"@ai-sdk/openai-compatible","env":["STACKIT_API_KEY"],"nm":8},"vercel":{"id":"vercel","name":"Vercel AI Gateway","doc":"https://github.com/vercel/ai/tree/5eb85cc45a259553501f535b8ac79a77d0e79223/packages/gateway","npm":"@ai-sdk/gateway","env":["AI_GATEWAY_API_KEY"],"nm":300},"submodel":{"id":"submodel","name":"submodel","doc":"https://submodel.gitbook.io","npm":"@ai-sdk/openai-compatible","env":["SUBMODEL_INSTAGEN_ACCESS_KEY"],"nm":9},"huggingface":{"id":"huggingface","name":"Hugging Face","doc":"https://huggingface.co/docs/inference-providers","npm":"@ai-sdk/openai-compatible","env":["HF_TOKEN"],"nm":51},"minimax-coding-plan":{"id":"minimax-coding-plan","name":"MiniMax Token Plan (minimax.io)","doc":"https://platform.minimax.io/docs/token-plan/intro","npm":"@ai-sdk/anthropic","env":["MINIMAX_API_KEY"],"nm":7},"novita-ai":{"id":"novita-ai","name":"NovitaAI","doc":"https://novita.ai/docs/guides/introduction","npm":"@ai-sdk/openai-compatible","env":["NOVITA_API_KEY"],"nm":105},"xai":{"id":"xai","name":"xAI","doc":"https://docs.x.ai/docs/models","npm":"@ai-sdk/xai","env":["XAI_API_KEY"],"nm":8},"privatemode-ai":{"id":"privatemode-ai","name":"Privatemode AI","doc":"https://docs.privatemode.ai/api/overview","npm":"@ai-sdk/openai-compatible","env":["PRIVATEMODE_API_KEY","PRIVATEMODE_ENDPOINT"],"nm":5},"drun":{"id":"drun","name":"D.Run (China)","doc":"https://www.d.run","npm":"@ai-sdk/openai-compatible","env":["DRUN_API_KEY"],"nm":3},"alibaba-token-plan-cn":{"id":"alibaba-token-plan-cn","name":"Alibaba Token Plan (China)","doc":"https://www.alibabacloud.com/help/zh/model-studio/token-plan-overview","npm":"@ai-sdk/openai-compatible","env":["ALIBABA_TOKEN_PLAN_API_KEY"],"nm":18},"moonshotai":{"id":"moonshotai","name":"Moonshot AI","doc":"https://platform.moonshot.ai/docs/api/chat","npm":"@ai-sdk/openai-compatible","env":["MOONSHOT_API_KEY"],"nm":9},"fireworks-ai":{"id":"fireworks-ai","name":"Fireworks AI","doc":"https://fireworks.ai/docs/","npm":"@ai-sdk/openai-compatible","env":["FIREWORKS_API_KEY"],"nm":16},"vultr":{"id":"vultr","name":"Vultr","doc":"https://api.vultrinference.com/","npm":"@ai-sdk/openai-compatible","env":["VULTR_API_KEY"],"nm":7},"302ai":{"id":"302ai","name":"302.AI","doc":"https://doc.302.ai","npm":"@ai-sdk/openai-compatible","env":["302AI_API_KEY"],"nm":97},"trustedrouter":{"id":"trustedrouter","name":"TrustedRouter","doc":"https://trustedrouter.com/docs","npm":"@ai-sdk/openai-compatible","env":["TRUSTEDROUTER_API_KEY"],"nm":7},"zhipuai":{"id":"zhipuai","name":"Zhipu AI","doc":"https://docs.z.ai/guides/overview/pricing","npm":"@ai-sdk/openai-compatible","env":["ZHIPU_API_KEY"],"nm":13},"cortecs":{"id":"cortecs","name":"Cortecs","doc":"https://api.cortecs.ai/v1/models","npm":"@ai-sdk/openai-compatible","env":["CORTECS_API_KEY"],"nm":56},"nebius":{"id":"nebius","name":"Nebius Token Factory","doc":"https://docs.tokenfactory.nebius.com/","npm":"@ai-sdk/openai-compatible","env":["NEBIUS_API_KEY"],"nm":30},"auriko":{"id":"auriko","name":"Auriko","doc":"https://docs.auriko.ai","npm":"@ai-sdk/openai-compatible","env":["AURIKO_API_KEY"],"nm":15},"stepfun-ai":{"id":"stepfun-ai","name":"StepFun AI","doc":"https://platform.stepfun.ai/docs/en/step-plan/integrations/open-code","npm":"@ai-sdk/openai-compatible","env":["STEPFUN_API_KEY"],"nm":8},"vivgrid":{"id":"vivgrid","name":"Vivgrid","doc":"https://docs.vivgrid.com/models","npm":"@ai-sdk/openai","env":["VIVGRID_API_KEY"],"nm":13},"tinfoil":{"id":"tinfoil","name":"Tinfoil","doc":"https://docs.tinfoil.sh","npm":"@ai-sdk/openai-compatible","env":["TINFOIL_API_KEY"],"nm":7},"mistral":{"id":"mistral","name":"Mistral","doc":"https://docs.mistral.ai/getting-started/models/","npm":"@ai-sdk/mistral","env":["MISTRAL_API_KEY"],"nm":30},"cloudflare-workers-ai":{"id":"cloudflare-workers-ai","name":"Cloudflare Workers AI","doc":"https://developers.cloudflare.com/workers-ai/models/","npm":"@ai-sdk/openai-compatible","env":["CLOUDFLARE_ACCOUNT_ID","CLOUDFLARE_API_KEY"],"nm":22},"bailing":{"id":"bailing","name":"Bailing","doc":"https://alipaytbox.yuque.com/sxs0ba/ling/intro","npm":"@ai-sdk/openai-compatible","env":["BAILING_API_TOKEN"],"nm":2},"anyapi":{"id":"anyapi","name":"AnyAPI","doc":"https://docs.anyapi.ai","npm":"@ai-sdk/openai-compatible","env":["ANYAPI_API_KEY"],"nm":30},"google":{"id":"google","name":"Google","doc":"https://ai.google.dev/gemini-api/docs/models","npm":"@ai-sdk/google","env":["GOOGLE_API_KEY","GOOGLE_GENERATIVE_AI_API_KEY","GEMINI_API_KEY"],"nm":22},"opencode-go":{"id":"opencode-go","name":"OpenCode Go","doc":"https://opencode.ai/docs/zen","npm":"@ai-sdk/openai-compatible","env":["OPENCODE_API_KEY"],"nm":19},"digitalocean":{"id":"digitalocean","name":"DigitalOcean","doc":"https://docs.digitalocean.com/products/gradient-ai-platform/details/models/","npm":"@ai-sdk/openai-compatible","env":["DIGITALOCEAN_ACCESS_TOKEN"],"nm":82},"subconscious":{"id":"subconscious","name":"Subconscious","doc":"https://docs.subconscious.dev","npm":"@ai-sdk/anthropic","env":["SUBCONSCIOUS_API_KEY"],"nm":2},"venice":{"id":"venice","name":"Venice AI","doc":"https://docs.venice.ai","npm":"venice-ai-sdk-provider","env":["VENICE_API_KEY"],"nm":75},"lmstudio":{"id":"lmstudio","name":"LMStudio","doc":"https://lmstudio.ai/models","npm":"@ai-sdk/openai-compatible","env":["LMSTUDIO_API_KEY"],"nm":3},"poolside":{"id":"poolside","name":"Poolside","doc":"https://platform.poolside.ai","npm":"@ai-sdk/openai-compatible","env":["POOLSIDE_API_KEY"],"nm":2},"zenmux":{"id":"zenmux","name":"ZenMux","doc":"https://docs.zenmux.ai","npm":"@ai-sdk/openai-compatible","env":["ZENMUX_API_KEY"],"nm":114},"kenari":{"id":"kenari","name":"Kenari","doc":"https://kenari.id/docs","npm":"@ai-sdk/openai-compatible","env":["KENARI_API_KEY"],"nm":23},"openai":{"id":"openai","name":"OpenAI","doc":"https://platform.openai.com/docs/models","npm":"@ai-sdk/openai","env":["OPENAI_API_KEY"],"nm":51},"berget":{"id":"berget","name":"Berget.AI","doc":"https://api.berget.ai","npm":"@ai-sdk/openai-compatible","env":["BERGET_API_KEY"],"nm":8},"snowflake-cortex":{"id":"snowflake-cortex","name":"Snowflake Cortex","doc":"https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api","npm":"@ai-sdk/openai-compatible","env":["SNOWFLAKE_ACCOUNT","SNOWFLAKE_CORTEX_PAT"],"nm":18},"tencent-token-plan":{"id":"tencent-token-plan","name":"Tencent Token Plan","doc":"https://cloud.tencent.com/document/product/1823/130060","npm":"@ai-sdk/openai-compatible","env":["TENCENT_TOKEN_PLAN_API_KEY"],"nm":1},"github-models":{"id":"github-models","name":"GitHub Models","doc":"https://docs.github.com/en/github-models","npm":"@ai-sdk/openai-compatible","env":["GITHUB_TOKEN"],"nm":55},"neuralwatt":{"id":"neuralwatt","name":"Neuralwatt","doc":"https://portal.neuralwatt.com/docs","npm":"@ai-sdk/openai-compatible","env":["NEURALWATT_API_KEY"],"nm":18},"siliconflow-cn":{"id":"siliconflow-cn","name":"SiliconFlow (China)","doc":"https://cloud.siliconflow.com/models","npm":"@ai-sdk/openai-compatible","env":["SILICONFLOW_CN_API_KEY"],"nm":47},"merge-gateway":{"id":"merge-gateway","name":"Merge Gateway","doc":"https://docs.merge.dev/merge-gateway","npm":"merge-gateway-ai-sdk-provider","env":["MERGE_GATEWAY_API_KEY"],"nm":93},"qihang-ai":{"id":"qihang-ai","name":"QiHang","doc":"https://www.qhaigc.net/docs","npm":"@ai-sdk/openai-compatible","env":["QIHANG_API_KEY"],"nm":9},"xiaomi-token-plan-ams":{"id":"xiaomi-token-plan-ams","name":"Xiaomi Token Plan (Europe)","doc":"https://platform.xiaomimimo.com/#/docs","npm":"@ai-sdk/openai-compatible","env":["XIAOMI_API_KEY"],"nm":8},"modelscope":{"id":"modelscope","name":"ModelScope","doc":"https://modelscope.cn/docs/model-service/API-Inference/intro","npm":"@ai-sdk/openai-compatible","env":["MODELSCOPE_API_KEY"],"nm":7},"groq":{"id":"groq","name":"Groq","doc":"https://console.groq.com/docs/models","npm":"@ai-sdk/groq","env":["GROQ_API_KEY"],"nm":15},"mixlayer":{"id":"mixlayer","name":"Mixlayer","doc":"https://docs.mixlayer.com","npm":"@ai-sdk/openai-compatible","env":["MIXLAYER_API_KEY"],"nm":5},"orcarouter":{"id":"orcarouter","name":"OrcaRouter","doc":"https://docs.orcarouter.ai","npm":"@ai-sdk/openai-compatible","env":["ORCAROUTER_API_KEY"],"nm":81},"helicone":{"id":"helicone","name":"Helicone","doc":"https://helicone.ai/models","npm":"@ai-sdk/openai-compatible","env":["HELICONE_API_KEY"],"nm":90},"zai":{"id":"zai","name":"Z.AI","doc":"https://docs.z.ai/guides/overview/pricing","npm":"@ai-sdk/openai-compatible","env":["ZHIPU_API_KEY"],"nm":14},"nearai":{"id":"nearai","name":"NEAR AI Cloud","doc":"https://docs.near.ai/","npm":"@ai-sdk/openai-compatible","env":["NEARAI_API_KEY"],"nm":37},"llmgateway":{"id":"llmgateway","name":"LLM Gateway","doc":"https://llmgateway.io/docs","npm":"@ai-sdk/openai-compatible","env":["LLMGATEWAY_API_KEY"],"nm":182},"alibaba-coding-plan-cn":{"id":"alibaba-coding-plan-cn","name":"Alibaba Coding Plan (China)","doc":"https://help.aliyun.com/zh/model-studio/coding-plan","npm":"@ai-sdk/openai-compatible","env":["ALIBABA_CODING_PLAN_API_KEY"],"nm":12},"abacus":{"id":"abacus","name":"Abacus","doc":"https://abacus.ai/help/api","npm":"@ai-sdk/openai-compatible","env":["ABACUS_API_KEY"],"nm":65},"cloudferro-sherlock":{"id":"cloudferro-sherlock","name":"CloudFerro Sherlock","doc":"https://docs.sherlock.cloudferro.com/","npm":"@ai-sdk/openai-compatible","env":["CLOUDFERRO_SHERLOCK_API_KEY"],"nm":5},"ollama-cloud":{"id":"ollama-cloud","name":"Ollama Cloud","doc":"https://docs.ollama.com/cloud","npm":"@ai-sdk/openai-compatible","env":["OLLAMA_API_KEY"],"nm":43},"cloudflare-ai-gateway":{"id":"cloudflare-ai-gateway","name":"Cloudflare AI Gateway","doc":"https://developers.cloudflare.com/ai-gateway/","npm":"ai-gateway-provider","env":["CLOUDFLARE_API_TOKEN","CLOUDFLARE_ACCOUNT_ID","CLOUDFLARE_GATEWAY_ID"],"nm":78},"moonshotai-cn":{"id":"moonshotai-cn","name":"Moonshot AI (China)","doc":"https://platform.moonshot.cn/docs/api/chat","npm":"@ai-sdk/openai-compatible","env":["MOONSHOT_API_KEY"],"nm":9},"morph":{"id":"morph","name":"Morph","doc":"https://docs.morphllm.com/api-reference/introduction","npm":"@ai-sdk/openai-compatible","env":["MORPH_API_KEY"],"nm":3},"sakana":{"id":"sakana","name":"Sakana AI","doc":"https://console.sakana.ai/models","npm":"@ai-sdk/openai-compatible","env":["SAKANA_API_KEY"],"nm":3},"deepinfra":{"id":"deepinfra","name":"Deep Infra","doc":"https://deepinfra.com/models","npm":"@ai-sdk/deepinfra","env":["DEEPINFRA_API_KEY"],"nm":40},"google-vertex-anthropic":{"id":"google-vertex-anthropic","name":"Vertex (Anthropic)","doc":"https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/claude","npm":"@ai-sdk/google-vertex/anthropic","env":["GOOGLE_VERTEX_PROJECT","GOOGLE_VERTEX_LOCATION","GOOGLE_APPLICATION_CREDENTIALS"],"nm":12},"v0":{"id":"v0","name":"v0","doc":"https://sdk.vercel.ai/providers/ai-sdk-providers/vercel","npm":"@ai-sdk/vercel","env":["V0_API_KEY"],"nm":3},"azure":{"id":"azure","name":"Azure","doc":"https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models","npm":"@ai-sdk/azure","env":["AZURE_RESOURCE_NAME","AZURE_API_KEY"],"nm":109},"cerebras":{"id":"cerebras","name":"Cerebras","doc":"https://inference-docs.cerebras.ai/models/overview","npm":"@ai-sdk/cerebras","env":["CEREBRAS_API_KEY"],"nm":3},"zai-coding-plan":{"id":"zai-coding-plan","name":"Z.AI Coding Plan","doc":"https://docs.z.ai/devpack/overview","npm":"@ai-sdk/openai-compatible","env":["ZHIPU_API_KEY"],"nm":6},"nvidia":{"id":"nvidia","name":"Nvidia","doc":"https://docs.api.nvidia.com/nim/","npm":"@ai-sdk/openai-compatible","env":["NVIDIA_API_KEY"],"nm":84},"evroc":{"id":"evroc","name":"evroc","doc":"https://docs.evroc.com/products/think/overview.html","npm":"@ai-sdk/openai-compatible","env":["EVROC_API_KEY"],"nm":15},"xiaomi":{"id":"xiaomi","name":"Xiaomi","doc":"https://platform.xiaomimimo.com/#/docs","npm":"@ai-sdk/openai-compatible","env":["XIAOMI_API_KEY"],"nm":6},"inception":{"id":"inception","name":"Inception","doc":"https://platform.inceptionlabs.ai/docs","npm":"@ai-sdk/openai-compatible","env":["INCEPTION_API_KEY"],"nm":2},"anthropic":{"id":"anthropic","name":"Anthropic","doc":"https://docs.anthropic.com/en/docs/about-claude/models","npm":"@ai-sdk/anthropic","env":["ANTHROPIC_API_KEY"],"nm":14},"tencent-coding-plan":{"id":"tencent-coding-plan","name":"Tencent Coding Plan (China)","doc":"https://cloud.tencent.com/document/product/1772/128947","npm":"@ai-sdk/openai-compatible","env":["TENCENT_CODING_PLAN_API_KEY"],"nm":8},"freemodel":{"id":"freemodel","name":"FreeModel","doc":"https://freemodel.dev","npm":"@ai-sdk/anthropic","env":["FREEMODEL_API_KEY"],"nm":10},"sap-ai-core":{"id":"sap-ai-core","name":"SAP AI Core","doc":"https://help.sap.com/docs/sap-ai-core","npm":"@jerome-benoit/sap-ai-provider-v2","env":["AICORE_SERVICE_KEY"],"nm":26},"opencode":{"id":"opencode","name":"OpenCode Zen","doc":"https://opencode.ai/docs/zen","npm":"@ai-sdk/openai-compatible","env":["OPENCODE_API_KEY"],"nm":74},"inference":{"id":"inference","name":"Inference","doc":"https://inference.net/models","npm":"@ai-sdk/openai-compatible","env":["INFERENCE_API_KEY"],"nm":9},"inceptron":{"id":"inceptron","name":"Inceptron","doc":"https://docs.inceptron.io","npm":"@ai-sdk/openai-compatible","env":["INCEPTRON_API_KEY"],"nm":6},"llama":{"id":"llama","name":"Llama","doc":"https://llama.developer.meta.com/docs/models","npm":"@ai-sdk/openai-compatible","env":["LLAMA_API_KEY"],"nm":7},"llmtr":{"id":"llmtr","name":"LLMTR","doc":"https://llmtr.com/docs","npm":"@ai-sdk/openai-compatible","env":["LLMTR_API_KEY"],"nm":6},"cohere":{"id":"cohere","name":"Cohere","doc":"https://docs.cohere.com/docs/models","npm":"@ai-sdk/cohere","env":["COHERE_API_KEY"],"nm":14},"sarvam":{"id":"sarvam","name":"Sarvam AI","doc":"https://docs.sarvam.ai/api-reference-docs/getting-started/models","npm":"@ai-sdk/openai-compatible","env":["SARVAM_API_KEY"],"nm":2},"stepfun":{"id":"stepfun","name":"StepFun","doc":"https://platform.stepfun.com/docs/zh/overview/concept","npm":"@ai-sdk/openai-compatible","env":["STEPFUN_API_KEY"],"nm":8},"hpc-ai":{"id":"hpc-ai","name":"HPC-AI","doc":"https://www.hpc-ai.com/doc/docs/quickstart/","npm":"@ai-sdk/openai-compatible","env":["HPC_AI_API_KEY"],"nm":3},"minimax-cn":{"id":"minimax-cn","name":"MiniMax (minimaxi.com)","doc":"https://platform.minimaxi.com/docs/guides/quickstart","npm":"@ai-sdk/anthropic","env":["MINIMAX_API_KEY"],"nm":7},"alibaba-coding-plan":{"id":"alibaba-coding-plan","name":"Alibaba Coding Plan","doc":"https://www.alibabacloud.com/help/en/model-studio/coding-plan","npm":"@ai-sdk/openai-compatible","env":["ALIBABA_CODING_PLAN_API_KEY"],"nm":12},"longcat":{"id":"longcat","name":"LongCat","doc":"https://longcat.chat/platform/docs/","npm":"@ai-sdk/openai-compatible","env":["LONGCAT_API_KEY"],"nm":1},"poe":{"id":"poe","name":"Poe","doc":"https://creator.poe.com/docs/external-applications/openai-compatible-api","npm":"@ai-sdk/openai-compatible","env":["POE_API_KEY"],"nm":137},"kimi-for-coding":{"id":"kimi-for-coding","name":"Kimi For Coding","doc":"https://www.kimi.com/code/docs/en/third-party-tools/other-coding-agents.html","npm":"@ai-sdk/anthropic","env":["KIMI_API_KEY"],"nm":4},"dinference":{"id":"dinference","name":"DInference","doc":"https://dinference.com","npm":"@ai-sdk/openai-compatible","env":["DINFERENCE_API_KEY"],"nm":5},"perplexity-agent":{"id":"perplexity-agent","name":"Perplexity Agent","doc":"https://docs.perplexity.ai/docs/agent-api/models","npm":"@ai-sdk/openai","env":["PERPLEXITY_API_KEY"],"nm":18},"siliconflow":{"id":"siliconflow","name":"SiliconFlow","doc":"https://cloud.siliconflow.com/models","npm":"@ai-sdk/openai-compatible","env":["SILICONFLOW_API_KEY"],"nm":49},"umans-ai-coding-plan":{"id":"umans-ai-coding-plan","name":"Umans AI Coding Plan","doc":"https://app.umans.ai/offers/code/docs","npm":"@ai-sdk/openai-compatible","env":["UMANS_AI_CODING_PLAN_API_KEY"],"nm":6},"io-net":{"id":"io-net","name":"IO.NET","doc":"https://io.net/docs/guides/intelligence/io-intelligence","npm":"@ai-sdk/openai-compatible","env":["IOINTELLIGENCE_API_KEY"],"nm":17},"gmicloud":{"id":"gmicloud","name":"GMI Cloud","doc":"https://docs.gmicloud.ai/inference-engine/api-reference/llm-api-reference","npm":"@ai-sdk/openai-compatible","env":["GMICLOUD_API_KEY"],"nm":13},"xiaomi-token-plan-cn":{"id":"xiaomi-token-plan-cn","name":"Xiaomi Token Plan (China)","doc":"https://platform.xiaomimimo.com/#/docs","npm":"@ai-sdk/openai-compatible","env":["XIAOMI_API_KEY"],"nm":8},"zeldoc":{"id":"zeldoc","name":"Zeldoc","doc":"https://docs.zeldoc.ai","npm":"@ai-sdk/openai-compatible","env":["ZELDOC_API_KEY"],"nm":1},"scaleway":{"id":"scaleway","name":"Scaleway","doc":"https://www.scaleway.com/en/docs/generative-apis/","npm":"@ai-sdk/openai-compatible","env":["SCALEWAY_API_KEY"],"nm":17},"ovhcloud":{"id":"ovhcloud","name":"OVHcloud AI Endpoints","doc":"https://www.ovhcloud.com/en/public-cloud/ai-endpoints/catalog//","npm":"@ai-sdk/openai-compatible","env":["OVHCLOUD_API_KEY"],"nm":14},"friendli":{"id":"friendli","name":"Friendli","doc":"https://friendli.ai/docs/guides/serverless_endpoints/introduction","npm":"@ai-sdk/openai-compatible","env":["FRIENDLI_TOKEN"],"nm":6},"tencent-tokenhub":{"id":"tencent-tokenhub","name":"Tencent TokenHub","doc":"https://cloud.tencent.com/document/product/1823/130050","npm":"@ai-sdk/openai-compatible","env":["TENCENT_TOKENHUB_API_KEY"],"nm":2},"wandb":{"id":"wandb","name":"Weights & Biases","doc":"https://docs.wandb.ai/guides/integrations/inference/","npm":"@ai-sdk/openai-compatible","env":["WANDB_API_KEY"],"nm":18},"kuae-cloud-coding-plan":{"id":"kuae-cloud-coding-plan","name":"KUAE Cloud Coding Plan","doc":"https://docs.mthreads.com/kuaecloud/kuaecloud-doc-online/coding_plan/","npm":"@ai-sdk/openai-compatible","env":["KUAE_API_KEY"],"nm":1},"gitlab":{"id":"gitlab","name":"GitLab Duo","doc":"https://docs.gitlab.com/user/duo_agent_platform/","npm":"gitlab-ai-provider","env":["GITLAB_TOKEN"],"nm":19},"kilo":{"id":"kilo","name":"Kilo Gateway","doc":"https://kilo.ai","npm":"@ai-sdk/openai-compatible","env":["KILO_API_KEY"],"nm":345},"lucidquery":{"id":"lucidquery","name":"LucidQuery","doc":"https://lucidquery.com/docs","npm":"@ai-sdk/openai-compatible","env":["LUCIDQUERY_API_KEY"],"nm":4},"meganova":{"id":"meganova","name":"Meganova","doc":"https://docs.meganova.ai","npm":"@ai-sdk/openai-compatible","env":["MEGANOVA_API_KEY"],"nm":19},"perplexity":{"id":"perplexity","name":"Perplexity","doc":"https://docs.perplexity.ai","npm":"@ai-sdk/perplexity","env":["PERPLEXITY_API_KEY"],"nm":4},"amazon-bedrock":{"id":"amazon-bedrock","name":"Amazon Bedrock","doc":"https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html","npm":"@ai-sdk/amazon-bedrock","env":["AWS_ACCESS_KEY_ID","AWS_SECRET_ACCESS_KEY","AWS_REGION","AWS_BEARER_TOKEN_BEDROCK"],"nm":107},"umans-ai":{"id":"umans-ai","name":"Umans AI","doc":"https://app.umans.ai/offers/code/docs/orgs","npm":"@ai-sdk/openai-compatible","env":["UMANS_AI_API_KEY"],"nm":5},"togetherai":{"id":"togetherai","name":"Together AI","doc":"https://docs.together.ai/docs/serverless-models","npm":"@ai-sdk/togetherai","env":["TOGETHER_API_KEY"],"nm":32},"frogbot":{"id":"frogbot","name":"FrogBot","doc":"https://docs.frogbot.ai","npm":"@ai-sdk/openai-compatible","env":["FROGBOT_API_KEY"],"nm":26},"openrouter":{"id":"openrouter","name":"OpenRouter","doc":"https://openrouter.ai/models","npm":"@openrouter/ai-sdk-provider","env":["OPENROUTER_API_KEY"],"nm":343},"jiekou":{"id":"jiekou","name":"Jiekou.AI","doc":"https://docs.jiekou.ai/docs/support/quickstart?utm_source=github_models.dev","npm":"@ai-sdk/openai-compatible","env":["JIEKOU_API_KEY"],"nm":61},"nova":{"id":"nova","name":"Nova","doc":"https://nova.amazon.com/dev/documentation","npm":"@ai-sdk/openai-compatible","env":["NOVA_API_KEY"],"nm":2},"alibaba-token-plan":{"id":"alibaba-token-plan","name":"Alibaba Token Plan","doc":"https://www.alibabacloud.com/help/en/model-studio/token-plan-overview","npm":"@ai-sdk/openai-compatible","env":["ALIBABA_TOKEN_PLAN_API_KEY"],"nm":18},"alibaba":{"id":"alibaba","name":"Alibaba","doc":"https://www.alibabacloud.com/help/en/model-studio/models","npm":"@ai-sdk/openai-compatible","env":["DASHSCOPE_API_KEY"],"nm":51},"databricks":{"id":"databricks","name":"Databricks","doc":"https://docs.databricks.com/aws/en/machine-learning/foundation-models/","npm":"@ai-sdk/openai-compatible","env":["DATABRICKS_HOST","DATABRICKS_TOKEN"],"nm":25},"crof":{"id":"crof","name":"CrofAI","doc":"https://crof.ai/docs","npm":"@ai-sdk/openai-compatible","env":["CROF_API_KEY"],"nm":23},"fastrouter":{"id":"fastrouter","name":"FastRouter","doc":"https://fastrouter.ai/models","npm":"@ai-sdk/openai-compatible","env":["FASTROUTER_API_KEY"],"nm":47},"abliteration-ai":{"id":"abliteration-ai","name":"abliteration.ai","doc":"https://docs.abliteration.ai/models","npm":"@ai-sdk/openai-compatible","env":["ABLIT_KEY"],"nm":1},"xpersona":{"id":"xpersona","name":"Xpersona","doc":"https://www.xpersona.co/docs","npm":"@ai-sdk/openai-compatible","env":["XPERSONA_API_KEY"],"nm":2},"azure-cognitive-services":{"id":"azure-cognitive-services","name":"Azure Cognitive Services","doc":"https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models","npm":"@ai-sdk/azure","env":["AZURE_COGNITIVE_SERVICES_RESOURCE_NAME","AZURE_COGNITIVE_SERVICES_API_KEY"],"nm":95},"baseten":{"id":"baseten","name":"Baseten","doc":"https://docs.baseten.co/inference/model-apis/overview","npm":"@ai-sdk/openai-compatible","env":["BASETEN_API_KEY"],"nm":13},"atomic-chat":{"id":"atomic-chat","name":"Atomic Chat","doc":"https://atomic.chat","npm":"@ai-sdk/openai-compatible","env":["ATOMIC_CHAT_API_KEY"],"nm":5},"routing-run":{"id":"routing-run","name":"routing.run","doc":"https://docs.routing.run/api-reference/models","npm":"@ai-sdk/openai-compatible","env":["ROUTING_RUN_API_KEY"],"nm":26},"aihubmix":{"id":"aihubmix","name":"AIHubMix","doc":"https://docs.aihubmix.com","npm":"@aihubmix/ai-sdk-provider","env":["AIHUBMIX_API_KEY"],"nm":55},"google-vertex":{"id":"google-vertex","name":"Vertex","doc":"https://cloud.google.com/vertex-ai/generative-ai/docs/models","npm":"@ai-sdk/google-vertex","env":["GOOGLE_VERTEX_PROJECT","GOOGLE_VERTEX_LOCATION","GOOGLE_APPLICATION_CREDENTIALS"],"nm":36},"nano-gpt":{"id":"nano-gpt","name":"NanoGPT","doc":"https://docs.nano-gpt.com","npm":"@ai-sdk/openai-compatible","env":["NANO_GPT_API_KEY"],"nm":617},"moark":{"id":"moark","name":"Moark","doc":"https://moark.com/docs/openapi/v1#tag/%E6%96%87%E6%9C%AC%E7%94%9F%E6%88%90","npm":"@ai-sdk/openai-compatible","env":["MOARK_API_KEY"],"nm":2},"lilac":{"id":"lilac","name":"Lilac","doc":"https://docs.getlilac.com/inference/models","npm":"@ai-sdk/openai-compatible","env":["LILAC_API_KEY"],"nm":4},"ambient":{"id":"ambient","name":"Ambient","doc":"https://ambient.xyz","npm":"@ai-sdk/openai-compatible","env":["AMBIENT_API_KEY"],"nm":3},"neon":{"id":"neon","name":"Neon","doc":"https://neon.com/docs","npm":"@ai-sdk/openai-compatible","env":["NEON_AI_GATEWAY_BASE_URL","NEON_AI_GATEWAY_TOKEN"],"nm":25},"upstage":{"id":"upstage","name":"Upstage","doc":"https://developers.upstage.ai/docs/apis/chat","npm":"@ai-sdk/openai-compatible","env":["UPSTAGE_API_KEY"],"nm":3},"zhipuai-coding-plan":{"id":"zhipuai-coding-plan","name":"Zhipu AI Coding Plan","doc":"https://docs.bigmodel.cn/cn/coding-plan/overview","npm":"@ai-sdk/openai-compatible","env":["ZHIPU_API_KEY"],"nm":7},"chutes":{"id":"chutes","name":"Chutes","doc":"https://llm.chutes.ai/v1/models","npm":"@ai-sdk/openai-compatible","env":["CHUTES_API_KEY"],"nm":13},"minimax-cn-coding-plan":{"id":"minimax-cn-coding-plan","name":"MiniMax Token Plan (minimaxi.com)","doc":"https://platform.minimaxi.com/docs/token-plan/intro","npm":"@ai-sdk/anthropic","env":["MINIMAX_API_KEY"],"nm":7},"deepseek":{"id":"deepseek","name":"DeepSeek","doc":"https://api-docs.deepseek.com/quick_start/pricing","npm":"@ai-sdk/openai-compatible","env":["DEEPSEEK_API_KEY"],"nm":4},"wafer.ai":{"id":"wafer.ai","name":"Wafer","doc":"https://docs.wafer.ai/wafer-pass","npm":"@ai-sdk/openai-compatible","env":["WAFER_API_KEY"],"nm":8},"minimax":{"id":"minimax","name":"MiniMax (minimax.io)","doc":"https://platform.minimax.io/docs/guides/quickstart","npm":"@ai-sdk/anthropic","env":["MINIMAX_API_KEY"],"nm":7},"github-copilot":{"id":"github-copilot","name":"GitHub Copilot","doc":"https://docs.github.com/en/copilot","npm":"@ai-sdk/openai-compatible","env":["GITHUB_TOKEN"],"nm":25},"clarifai":{"id":"clarifai","name":"Clarifai","doc":"https://docs.clarifai.com/compute/inference/","npm":"@ai-sdk/openai-compatible","env":["CLARIFAI_PAT"],"nm":12},"the-grid-ai":{"id":"the-grid-ai","name":"The Grid AI","doc":"https://thegrid.ai/docs","npm":"@ai-sdk/openai-compatible","env":["THEGRIDAI_API_KEY"],"nm":9},"synthetic":{"id":"synthetic","name":"Synthetic","doc":"https://synthetic.new/pricing","npm":"@ai-sdk/openai-compatible","env":["SYNTHETIC_API_KEY"],"nm":7},"iflowcn":{"id":"iflowcn","name":"iFlow","doc":"https://platform.iflow.cn/en/docs","npm":"@ai-sdk/openai-compatible","env":["IFLOW_API_KEY"],"nm":14},"xiaomi-token-plan-sgp":{"id":"xiaomi-token-plan-sgp","name":"Xiaomi Token Plan (Singapore)","doc":"https://platform.xiaomimimo.com/#/docs","npm":"@ai-sdk/openai-compatible","env":["XIAOMI_API_KEY"],"nm":8},"claudinio":{"id":"claudinio","name":"Claudinio","doc":"https://claudin.io","npm":"@ai-sdk/openai-compatible","env":["CLAUDINIO_API_KEY"],"nm":2}},"models":[{"p":"requesty","id":"xai/grok-4","n":"Grok 4","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"cw":3,"ctx":256000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-09","lu":"2025-09-09"},{"p":"requesty","id":"xai/grok-4-fast","n":"Grok 4 Fast","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"cw":0.2,"ctx":2000000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-19","lu":"2025-09-19"},{"p":"requesty","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.31,"cw":2.375,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"requesty","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"cw":0.55,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"requesty","id":"google/gemini-3-pro-preview","n":"Gemini 3 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"cw":4.5,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"requesty","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"cw":1,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"requesty","id":"openai/gpt-5.2-chat","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"requesty","id":"openai/gpt-5.2-pro","n":"GPT-5.2 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"requesty","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"taiv","mo":"tai","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"requesty","id":"openai/gpt-5-chat","n":"GPT-5 Chat (latest)","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"requesty","id":"openai/gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"requesty","id":"openai/o4-mini","n":"o4 Mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.28,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2025-04-16","lu":"2025-04-16"},{"p":"requesty","id":"openai/gpt-5.1-chat","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"requesty","id":"openai/gpt-5.1-codex","n":"GPT-5.1-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"requesty","id":"openai/gpt-5.1-codex-max","n":"GPT-5.1-Codex-Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.1,"co":9,"cr":0.11,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"requesty","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"requesty","id":"openai/gpt-5.3-codex","n":"GPT-5.3-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-24","lu":"2026-02-24"},{"p":"requesty","id":"openai/gpt-5.1-codex-mini","n":"GPT-5.1-Codex-Mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"requesty","id":"openai/gpt-5-image","n":"GPT-5 Image","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":10,"cr":1.25,"ctx":400000,"out":128000,"mi":"tip","mo":"ti","rs":1,"tc":1,"kn":"2024-10-01","rd":"2025-10-14","lu":"2025-10-14"},{"p":"requesty","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"requesty","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"requesty","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.03,"ctx":128000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"requesty","id":"openai/gpt-4.1-mini","n":"GPT-4.1 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"requesty","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.01,"ctx":16000,"out":4000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"requesty","id":"openai/gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"cr":30,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"requesty","id":"openai/gpt-4o-mini","n":"GPT-4o Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"cr":0.08,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2024-07-18","lu":"2024-07-18"},{"p":"requesty","id":"openai/gpt-5-codex","n":"GPT-5 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10-01","rd":"2025-09-15","lu":"2025-09-15"},{"p":"requesty","id":"openai/gpt-5.2-codex","n":"GPT-5.2-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-14","lu":"2026-01-14"},{"p":"requesty","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"requesty","id":"anthropic/claude-3-7-sonnet","n":"Claude Sonnet 3.7","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-01","rd":"2025-02-19","lu":"2025-02-19"},{"p":"requesty","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"requesty","id":"anthropic/claude-opus-4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"requesty","id":"anthropic/claude-sonnet-4-5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"requesty","id":"anthropic/claude-opus-4-1","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"requesty","id":"anthropic/claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":62000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-01","rd":"2025-10-15","lu":"2025-10-15"},{"p":"requesty","id":"anthropic/claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"requesty","id":"anthropic/claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"requesty","id":"anthropic/claude-opus-4","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"qiniu-ai","id":"deepseek-r1-0528","n":"DeepSeek-R1-0528","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"doubao-1.5-thinking-pro","n":"Doubao 1.5 Thinking Pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":128000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-vl-30b-a3b-thinking","n":"Qwen3-Vl 30b A3b Thinking","d":"Multimodal model for analyzing text, images, documents, and rich media","ctx":128000,"out":32000,"mi":"tiv","mo":"t","tc":1,"rd":"2026-02-09","lu":"2026-02-09"},{"p":"qiniu-ai","id":"claude-3.5-haiku","n":"Claude 3.5 Haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2025-08-26","lu":"2025-08-26"},{"p":"qiniu-ai","id":"deepseek-v3-0324","n":"DeepSeek-V3-0324","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":16000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-235b-a22b-instruct-2507","n":"Qwen3 235b A22B Instruct 2507","d":"Tool-capable chat model for instruction following and agentic application workflows","ctx":262144,"out":64000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-12","lu":"2025-08-12"},{"p":"qiniu-ai","id":"deepseek-v3","n":"DeepSeek-V3","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":16000,"mi":"t","mo":"t","rd":"2025-08-13","lu":"2025-08-13"},{"p":"qiniu-ai","id":"kimi-k2","n":"Kimi K2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-32b","n":"Qwen3 32B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":40000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-max-preview","n":"Qwen3 Max Preview","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":256000,"out":64000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-06","lu":"2025-09-06"},{"p":"qiniu-ai","id":"claude-3.5-sonnet","n":"Claude 3.5 Sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ctx":200000,"out":8200,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-09-09","lu":"2025-09-09"},{"p":"qiniu-ai","id":"qwen3-next-80b-a3b-instruct","n":"Qwen3 Next 80B A3B Instruct","d":"Tool-capable chat model for instruction following and agentic application workflows","ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2025-09-12","lu":"2025-09-12"},{"p":"qiniu-ai","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ctx":1048576,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"claude-4.5-haiku","n":"Claude 4.5 Haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-16","lu":"2025-10-16"},{"p":"qiniu-ai","id":"kling-v2-6","n":"Kling-V2 6","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":99999999,"out":99999999,"mi":"tiv","mo":"v","rd":"2026-01-13","lu":"2026-01-13"},{"p":"qiniu-ai","id":"glm-4.5","n":"GLM 4.5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"claude-4.1-opus","n":"Claude 4.1 Opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"qiniu-ai","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ctx":1048576,"out":64000,"mi":"tiav","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-max","n":"Qwen3 Max","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-09-24","lu":"2025-09-24"},{"p":"qiniu-ai","id":"doubao-seed-2.0-pro","n":"Doubao Seed 2.0 Pro","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"qiniu-ai","id":"doubao-seed-1.6","n":"Doubao-Seed 1.6","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":32000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2025-08-15","lu":"2025-08-15"},{"p":"qiniu-ai","id":"doubao-seed-1.6-thinking","n":"Doubao-Seed 1.6 Thinking","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":32000,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2025-08-15","lu":"2025-08-15"},{"p":"qiniu-ai","id":"gemini-2.0-flash","n":"Gemini 2.0 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ctx":1048576,"out":8192,"mi":"tiav","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen-max-2025-01-25","n":"Qwen2.5-Max-2025-01-25","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"claude-4.0-sonnet","n":"Claude 4.0 Sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"doubao-1.5-pro-32k","n":"Doubao 1.5 Pro 32k","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":128000,"out":12000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-30b-a3b-instruct-2507","n":"Qwen3 30b A3b Instruct 2507","d":"Tool-capable chat model for instruction following and agentic application workflows","ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"rd":"2026-02-04","lu":"2026-02-04"},{"p":"qiniu-ai","id":"qwen3-next-80b-a3b-thinking","n":"Qwen3 Next 80B A3B Thinking","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-12","lu":"2025-09-12"},{"p":"qiniu-ai","id":"qwen3-235b-a22b-thinking-2507","n":"Qwen3 235B A22B Thinking 2507","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":262144,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-12","lu":"2025-08-12"},{"p":"qiniu-ai","id":"deepseek-r1","n":"DeepSeek-R1","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"doubao-1.5-vision-pro","n":"Doubao 1.5 Vision Pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":128000,"out":16000,"mi":"tiv","mo":"t","rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"gemini-3.0-pro-image-preview","n":"Gemini 3.0 Pro Image Preview","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":32768,"out":8192,"mi":"ti","mo":"ti","rd":"2025-11-20","lu":"2025-11-20"},{"p":"qiniu-ai","id":"gemini-2.5-flash-image","n":"Gemini 2.5 Flash Image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":32768,"out":8192,"mi":"ti","mo":"i","rd":"2025-10-22","lu":"2025-10-22"},{"p":"qiniu-ai","id":"gemini-2.5-flash-lite","n":"Gemini 2.5 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ctx":1048576,"out":64000,"mi":"tiav","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"claude-3.7-sonnet","n":"Claude 3.7 Sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ctx":200000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-30b-a3b-thinking-2507","n":"Qwen3 30b A3b Thinking 2507","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":126000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-04","lu":"2026-02-04"},{"p":"qiniu-ai","id":"qwen2.5-vl-72b-instruct","n":"Qwen 2.5 VL 72B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ctx":128000,"out":8192,"mi":"tiav","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"gpt-oss-120b","n":"gpt-oss-120b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"qiniu-ai","id":"doubao-seed-1.6-flash","n":"Doubao-Seed 1.6 Flash","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":32000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2025-08-15","lu":"2025-08-15"},{"p":"qiniu-ai","id":"deepseek-v3.1","n":"DeepSeek-V3.1","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-19","lu":"2025-08-19"},{"p":"qiniu-ai","id":"qwen3-235b-a22b","n":"Qwen 3 235B A22B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-coder-480b-a35b-instruct","n":"Qwen3 Coder 480B A35B Instruct","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":262000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2025-08-14","lu":"2025-08-14"},{"p":"qiniu-ai","id":"qwen3.5-397b-a17b","n":"Qwen3.5 397B A17B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-22","lu":"2026-02-22"},{"p":"qiniu-ai","id":"mimo-v2-flash","n":"Mimo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"cr":0.01,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12-01","rd":"2025-12-16","lu":"2026-02-04"},{"p":"qiniu-ai","id":"qwen-vl-max-2025-01-25","n":"Qwen VL-MAX-2025-01-25","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ctx":128000,"out":4096,"mi":"tiav","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen2.5-vl-7b-instruct","n":"Qwen 2.5 VL 7B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ctx":128000,"out":8192,"mi":"tiav","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"glm-4.5-air","n":"GLM 4.5 Air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ctx":131000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"claude-4.5-opus","n":"Claude 4.5 Opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ctx":200000,"out":200000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-25","lu":"2025-11-25"},{"p":"qiniu-ai","id":"gemini-3.0-pro-preview","n":"Gemini 3.0 Pro Preview","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ctx":1000000,"out":64000,"mi":"tivpa","mo":"t","rs":1,"tc":1,"rd":"2025-11-19","lu":"2025-11-19"},{"p":"qiniu-ai","id":"doubao-seed-2.0-mini","n":"Doubao Seed 2.0 Mini","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":32000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"qiniu-ai","id":"claude-4.0-opus","n":"Claude 4.0 Opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"gpt-oss-20b","n":"gpt-oss-20b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"qiniu-ai","id":"gemini-3.0-flash-preview","n":"Gemini 3.0 Flash Preview","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ctx":1000000,"out":64000,"mi":"tiavp","mo":"t","rs":1,"tc":1,"rd":"2025-12-18","lu":"2025-12-18"},{"p":"qiniu-ai","id":"MiniMax-M1","n":"MiniMax M1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":1000000,"out":80000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen-turbo","n":"Qwen-Turbo","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ctx":1000000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"qwen3-30b-a3b","n":"Qwen3 30B A3B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":40000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"claude-4.5-sonnet","n":"Claude 4.5 Sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-09-30","lu":"2025-09-30"},{"p":"qiniu-ai","id":"doubao-seed-2.0-lite","n":"Doubao Seed 2.0 Lite","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":256000,"out":32000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"qiniu-ai","id":"gemini-2.0-flash-lite","n":"Gemini 2.0 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ctx":1048576,"out":8192,"mi":"tiav","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"qiniu-ai","id":"doubao-seed-2.0-code","n":"Doubao Seed 2.0 Code","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":256000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"qiniu-ai","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ctx":256000,"out":100000,"mi":"t","mo":"t","tc":1,"rd":"2025-11-07","lu":"2025-11-07"},{"p":"qiniu-ai","id":"moonshotai/kimi-k2.5","n":"Moonshotai/Kimi-K2.5","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ctx":256000,"out":256000,"mi":"tiv","mo":"t","tc":1,"rd":"2026-01-28","lu":"2026-01-28"},{"p":"qiniu-ai","id":"moonshotai/kimi-k2-0905","n":"Kimi K2 0905","d":"Kimi model for long-context chat, coding, and agentic reasoning","ctx":256000,"out":100000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-08","lu":"2025-09-08"},{"p":"qiniu-ai","id":"stepfun-ai/gelab-zero-4b-preview","n":"Stepfun-Ai/Gelab Zero 4b Preview","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ctx":8192,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"qiniu-ai","id":"x-ai/grok-code-fast-1","n":"x-AI/Grok-Code-Fast 1","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":256000,"out":10000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-02","lu":"2025-09-02"},{"p":"qiniu-ai","id":"x-ai/grok-4.1-fast-reasoning","n":"X-Ai/Grok 4.1 Fast Reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":20000000,"out":2000000,"mi":"tiav","mo":"t","rs":1,"tc":1,"rd":"2025-12-19","lu":"2025-12-19"},{"p":"qiniu-ai","id":"x-ai/grok-4.1-fast-non-reasoning","n":"X-Ai/Grok 4.1 Fast Non Reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":2000000,"mi":"tiav","mo":"t","tc":1,"rd":"2025-12-19","lu":"2025-12-19"},{"p":"qiniu-ai","id":"x-ai/grok-4-fast-reasoning","n":"X-Ai/Grok-4-Fast-Reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":2000000,"mi":"tiav","mo":"t","rs":1,"tc":1,"rd":"2025-12-18","lu":"2025-12-18"},{"p":"qiniu-ai","id":"x-ai/grok-4-fast","n":"x-AI/Grok-4-Fast","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":2000000,"mi":"tiav","mo":"t","rs":1,"tc":1,"rd":"2025-09-20","lu":"2025-09-20"},{"p":"qiniu-ai","id":"x-ai/grok-4-fast-non-reasoning","n":"X-Ai/Grok-4-Fast-Non-Reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":2000000,"mi":"tiav","mo":"t","tc":1,"rd":"2025-12-18","lu":"2025-12-18"},{"p":"qiniu-ai","id":"x-ai/grok-4.1-fast","n":"x-AI/Grok-4.1-Fast","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":2000000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-11-20","lu":"2025-11-20"},{"p":"qiniu-ai","id":"z-ai/autoglm-phone-9b","n":"Z-Ai/Autoglm Phone 9b","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ctx":12800,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"qiniu-ai","id":"z-ai/glm-4.7","n":"Z-Ai/GLM 4.7","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":200000,"out":200000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"qiniu-ai","id":"z-ai/glm-4.6","n":"Z-AI/GLM 4.6","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":200000,"out":200000,"mi":"t","mo":"t","tc":1,"rd":"2025-10-11","lu":"2025-10-11"},{"p":"qiniu-ai","id":"z-ai/glm-5","n":"Z-Ai/GLM 5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"qiniu-ai","id":"openai/gpt-5","n":"OpenAI/GPT-5","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ctx":400000,"out":128000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-19","lu":"2025-09-19"},{"p":"qiniu-ai","id":"openai/gpt-5.2","n":"OpenAI/GPT-5.2","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-11","lu":"2025-12-11"},{"p":"qiniu-ai","id":"xiaomi/mimo-v2-flash","n":"Xiaomi/Mimo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"cr":0.01,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12-01","rd":"2025-12-16","lu":"2026-02-04"},{"p":"qiniu-ai","id":"stepfun/step-3.5-flash","n":"Stepfun/Step-3.5 Flash","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ctx":64000,"out":4096,"mi":"ti","mo":"t","rd":"2026-02-02","lu":"2026-02-02"},{"p":"qiniu-ai","id":"meituan/longcat-flash-lite","n":"Meituan/Longcat-Flash-Lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":256000,"out":320000,"mi":"t","mo":"t","tc":1,"rd":"2026-02-06","lu":"2026-02-06"},{"p":"qiniu-ai","id":"meituan/longcat-flash-chat","n":"Meituan/Longcat-Flash-Chat","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":131072,"out":131072,"mi":"t","mo":"t","rd":"2025-11-05","lu":"2025-11-05"},{"p":"qiniu-ai","id":"deepseek/deepseek-v3.1-terminus","n":"DeepSeek/DeepSeek-V3.1-Terminus","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-22","lu":"2025-09-22"},{"p":"qiniu-ai","id":"deepseek/deepseek-v3.1-terminus-thinking","n":"DeepSeek/DeepSeek-V3.1-Terminus-Thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"rd":"2025-09-22","lu":"2025-09-22"},{"p":"qiniu-ai","id":"deepseek/deepseek-v3.2-exp-thinking","n":"DeepSeek/DeepSeek-V3.2-Exp-Thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"qiniu-ai","id":"deepseek/deepseek-v3.2-exp","n":"DeepSeek/DeepSeek-V3.2-Exp","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"qiniu-ai","id":"deepseek/deepseek-v3.2-251201","n":"Deepseek/DeepSeek-V3.2","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"qiniu-ai","id":"deepseek/deepseek-math-v2","n":"Deepseek/Deepseek-Math-V2","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":160000,"out":160000,"mi":"t","mo":"t","rs":1,"rd":"2025-12-04","lu":"2025-12-04"},{"p":"qiniu-ai","id":"minimax/minimax-m2.5","n":"Minimax/Minimax-M2.5","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"qiniu-ai","id":"minimax/minimax-m2.1","n":"Minimax/Minimax-M2.1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"qiniu-ai","id":"minimax/minimax-m2","n":"Minimax/Minimax-M2","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-10-28","lu":"2025-10-28"},{"p":"qiniu-ai","id":"minimax/minimax-m2.5-highspeed","n":"Minimax/Minimax-M2.5 Highspeed","d":"High-speed MiniMax model for low-latency coding and agent workflows","ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"alibaba-cn","id":"qwen2-5-math-72b-instruct","n":"Qwen2.5-Math 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.574,"co":1.721,"ctx":4096,"out":3072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"deepseek-r1-0528","n":"DeepSeek R1 0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.574,"co":2.294,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-28","lu":"2025-05-28"},{"p":"alibaba-cn","id":"qwen3-omni-flash","n":"Qwen3-Omni Flash","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.058,"co":0.23,"ctx":65536,"out":16384,"mi":"tiav","mo":"ta","rs":1,"tc":1,"kn":"2024-04","rd":"2025-09-15","lu":"2025-09-15"},{"p":"alibaba-cn","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"alibaba-cn","id":"qwen-plus","n":"Qwen Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.115,"co":0.287,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-09-11"},{"p":"alibaba-cn","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3-Coder 30B-A3B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.216,"co":0.861,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwen2-5-coder-7b-instruct","n":"Qwen2.5-Coder 7B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.144,"co":0.287,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-11","lu":"2024-11"},{"p":"alibaba-cn","id":"deepseek-v3","n":"DeepSeek V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.287,"co":1.147,"ctx":65536,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"alibaba-cn","id":"qwen3-omni-flash-realtime","n":"Qwen3-Omni Flash Realtime","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.23,"co":0.918,"ctx":65536,"out":16384,"mi":"tia","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-09-15","lu":"2025-09-15"},{"p":"alibaba-cn","id":"deepseek-r1-distill-llama-70b","n":"DeepSeek R1 Distill Llama 70B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.287,"co":0.861,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"qwen3-32b","n":"Qwen3 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.287,"co":1.147,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwen-omni-turbo-realtime","n":"Qwen-Omni Turbo Realtime","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.23,"co":0.918,"ctx":32768,"out":2048,"mi":"tia","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-05-08","lu":"2025-05-08"},{"p":"alibaba-cn","id":"qwen2-5-math-7b-instruct","n":"Qwen2.5-Math 7B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.144,"co":0.287,"ctx":4096,"out":3072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"qwen3-next-80b-a3b-instruct","n":"Qwen3-Next 80B-A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.144,"co":0.574,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"alibaba-cn","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"alibaba-cn","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.5,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"alibaba-cn","id":"qwen-long","n":"Qwen Long","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.072,"co":0.287,"ctx":10000000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2025-01-25","lu":"2025-01-25"},{"p":"alibaba-cn","id":"qwen-math-turbo","n":"Qwen Math Turbo","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.287,"co":0.861,"ctx":4096,"out":3072,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-09-19","lu":"2024-09-19"},{"p":"alibaba-cn","id":"qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0.861,"co":3.441,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"alibaba-cn","id":"qwen2-5-omni-7b","n":"Qwen2.5-Omni 7B","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.087,"co":0.345,"ctx":32768,"out":2048,"mi":"tiav","mo":"ta","tc":1,"ow":1,"kn":"2024-04","rd":"2024-12","lu":"2024-12"},{"p":"alibaba-cn","id":"qwen3-8b","n":"Qwen3 8B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.072,"co":0.287,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwen2-5-14b-instruct","n":"Qwen2.5 14B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.144,"co":0.431,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.87,"co":3.48,"cr":0.17,"ctx":202752,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-14","lu":"2026-04-14"},{"p":"alibaba-cn","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"alibaba-cn","id":"qwen3-next-80b-a3b-thinking","n":"Qwen3-Next 80B-A3B (Thinking)","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.144,"co":1.434,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"alibaba-cn","id":"qvq-max","n":"QVQ Max","f":"qvq","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":1.147,"co":4.588,"ctx":131072,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-03-25","lu":"2025-03-25"},{"p":"alibaba-cn","id":"qwen-plus-character","n":"Qwen Plus Character","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.115,"co":0.287,"ctx":32768,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01","lu":"2024-01"},{"p":"alibaba-cn","id":"kimi-k2-thinking","n":"Moonshot Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.574,"co":2.294,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-06","lu":"2025-11-06"},{"p":"alibaba-cn","id":"deepseek-r1","n":"DeepSeek R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.574,"co":2.294,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"qwen3.5-flash","n":"Qwen3.5 Flash","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.172,"co":1.72,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-23","lu":"2026-02-23"},{"p":"alibaba-cn","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.1,"co":3.851,"cr":0.275,"cw":0,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"alibaba-cn","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1875,"co":1.125,"cw":0.234375,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"alibaba-cn","id":"qwen2-5-vl-72b-instruct","n":"Qwen2.5-VL 72B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":2.294,"co":6.881,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"qwen3-vl-plus","n":"Qwen3-VL Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.143353,"co":1.433525,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"alibaba-cn","id":"qwen-vl-ocr","n":"Qwen-VL OCR","f":"qwen","d":"OCR model for extracting structured text from documents and screenshots","ci":0.717,"co":0.717,"ctx":34096,"out":4096,"mi":"ti","mo":"t","kn":"2024-04","rd":"2024-10-28","lu":"2025-04-13"},{"p":"alibaba-cn","id":"qwen-mt-turbo","n":"Qwen-MT Turbo","f":"qwen","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":0.101,"co":0.28,"ctx":16384,"out":8192,"mi":"t","mo":"t","kn":"2024-04","rd":"2025-01","lu":"2025-01"},{"p":"alibaba-cn","id":"qwen-math-plus","n":"Qwen Math Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.574,"co":1.721,"ctx":4096,"out":3072,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-08-16","lu":"2024-09-19"},{"p":"alibaba-cn","id":"qwen-mt-plus","n":"Qwen-MT Plus","f":"qwen","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":0.259,"co":0.775,"ctx":16384,"out":8192,"mi":"t","mo":"t","kn":"2024-04","rd":"2025-01","lu":"2025-01"},{"p":"alibaba-cn","id":"qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.573,"co":3.44,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"alibaba-cn","id":"qwen-omni-turbo","n":"Qwen-Omni Turbo","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.058,"co":0.23,"ctx":32768,"out":2048,"mi":"tiav","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-01-19","lu":"2025-03-26"},{"p":"alibaba-cn","id":"qwen2-5-72b-instruct","n":"Qwen2.5 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.574,"co":1.721,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"deepseek-r1-distill-qwen-7b","n":"DeepSeek R1 Distill Qwen 7B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.072,"co":0.144,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"deepseek-v3-1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.574,"co":1.721,"ctx":131072,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"qwen2-5-coder-32b-instruct","n":"Qwen2.5-Coder 32B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.287,"co":0.861,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-11","lu":"2024-11"},{"p":"alibaba-cn","id":"qwen-flash","n":"Qwen Flash","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.022,"co":0.216,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"alibaba-cn","id":"deepseek-v3-2-exp","n":"DeepSeek V3.2 Exp","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.287,"co":0.431,"ctx":131072,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"moonshot-kimi-k2-instruct","n":"Moonshot Kimi K2 Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.574,"co":2.294,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"deepseek-r1-distill-qwen-14b","n":"DeepSeek R1 Distill Qwen 14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.144,"co":0.431,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"qwen3-vl-235b-a22b","n":"Qwen3-VL 235B-A22B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.286705,"co":1.14682,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwen-deep-research","n":"Qwen Deep Research","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":7.742,"co":23.367,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01","lu":"2024-01"},{"p":"alibaba-cn","id":"kimi-k2.5","n":"Moonshot Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.574,"co":2.411,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"alibaba-cn","id":"qwen3-vl-30b-a3b","n":"Qwen3-VL 30B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.108,"co":0.431,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwen-vl-max","n":"Qwen-VL Max","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.23,"co":0.574,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-04-08","lu":"2025-08-13"},{"p":"alibaba-cn","id":"deepseek-r1-distill-qwen-1-5b","n":"DeepSeek R1 Distill Qwen 1.5B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"qwen-max","n":"Qwen Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0.345,"co":1.377,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-04-03","lu":"2025-01-25"},{"p":"alibaba-cn","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-cn","id":"qwen3-235b-a22b","n":"Qwen3 235B-A22B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.287,"co":1.147,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"deepseek-r1-distill-llama-8b","n":"DeepSeek R1 Distill Llama 8B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0,"co":0,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"kimi-k2.6","n":"Moonshot Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.929,"co":3.858,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"alibaba-cn","id":"qwen3-coder-480b-a35b-instruct","n":"Qwen3-Coder 480B-A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.861,"co":3.441,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwq-plus","n":"QwQ Plus","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.23,"co":0.574,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-03-05","lu":"2025-03-05"},{"p":"alibaba-cn","id":"qwen2-5-32b-instruct","n":"Qwen2.5 32B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.287,"co":0.861,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"qwen3.5-397b-a17b","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.43,"co":2.58,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"alibaba-cn","id":"tongyi-intent-detect-v3","n":"Tongyi Intent Detect V3","f":"yi","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.058,"co":0.144,"ctx":8192,"out":1024,"mi":"t","mo":"t","kn":"2024-04","rd":"2024-01","lu":"2024-01"},{"p":"alibaba-cn","id":"qwen3-coder-flash","n":"Qwen3 Coder Flash","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.144,"co":0.574,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"alibaba-cn","id":"deepseek-r1-distill-qwen-32b","n":"DeepSeek R1 Distill Qwen 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.287,"co":0.861,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-01","lu":"2025-01-01"},{"p":"alibaba-cn","id":"qwq-32b","n":"QwQ 32B","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.287,"co":0.861,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-04","rd":"2024-12","lu":"2024-12"},{"p":"alibaba-cn","id":"qwen3-14b","n":"Qwen3 14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.144,"co":0.574,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba-cn","id":"qwen3-asr-flash","n":"Qwen3-ASR Flash","f":"qwen","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0.032,"co":0.032,"ctx":53248,"out":4096,"mi":"a","mo":"t","kn":"2024-04","rd":"2025-09-08","lu":"2025-09-08"},{"p":"alibaba-cn","id":"qwen-doc-turbo","n":"Qwen Doc Turbo","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.087,"co":0.144,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01","lu":"2024-01"},{"p":"alibaba-cn","id":"glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.86,"co":3.15,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"alibaba-cn","id":"qwen-turbo","n":"Qwen Turbo","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.044,"co":0.087,"ctx":1000000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2024-11-01","lu":"2025-07-15"},{"p":"alibaba-cn","id":"qwen2-5-7b-instruct","n":"Qwen2.5 7B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.072,"co":0.144,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"qwen2-5-vl-7b-instruct","n":"Qwen2.5-VL 7B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.287,"co":0.717,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba-cn","id":"qwen3.6-max-preview","n":"Qwen3.6 Max Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.32,"co":7.9,"cr":0.132,"ctx":245800,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-20","lu":"2026-04-21"},{"p":"alibaba-cn","id":"qwen-vl-plus","n":"Qwen-VL Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.115,"co":0.287,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-08-15"},{"p":"alibaba-cn","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"alibaba-cn","id":"MiniMax/MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"alibaba-cn","id":"kimi/kimi-k2.5","n":"kimi/kimi-k2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"alibaba-cn","id":"siliconflow/deepseek-r1-0528","n":"siliconflow/deepseek-r1-0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.18,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-28","lu":"2025-11-25"},{"p":"alibaba-cn","id":"siliconflow/deepseek-v3.1-terminus","n":"siliconflow/deepseek-v3.1-terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2025-11-25"},{"p":"alibaba-cn","id":"siliconflow/deepseek-v3-0324","n":"siliconflow/deepseek-v3-0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.25,"co":1,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"rd":"2024-12-26","lu":"2025-11-25"},{"p":"alibaba-cn","id":"siliconflow/deepseek-v3.2","n":"siliconflow/deepseek-v3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.42,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-03","lu":"2025-12-03"},{"p":"alibaba-cn","id":"qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Hosted Qwen coder for software agents, repo edits, and long-context code","ci":1,"co":5,"ctx":1048576,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"regolo-ai","id":"llama-3.1-8b-instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.05,"co":0.25,"ctx":120000,"out":120000,"mi":"t","mo":"t","tc":1,"rd":"2025-04-07","lu":"2025-04-07"},{"p":"regolo-ai","id":"minimax-m2.5","n":"MiniMax 2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.8,"co":3.5,"ctx":190000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-10","lu":"2026-03-10"},{"p":"regolo-ai","id":"mistral-small3.2","n":"Mistral Small 3.2","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.5,"co":2.2,"ctx":120000,"out":120000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-31","lu":"2025-01-31"},{"p":"regolo-ai","id":"qwen3-reranker-4b","n":"Qwen3-Reranker-4B","f":"qwen","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0.12,"co":0.12,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"regolo-ai","id":"qwen3-embedding-8b","n":"Qwen3-Embedding-8B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0.1,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"regolo-ai","id":"llama-3.3-70b-instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.6,"co":2.7,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-04-28","lu":"2025-04-28"},{"p":"regolo-ai","id":"qwen-image","n":"Qwen-Image","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":2,"ctx":8192,"out":4096,"mi":"t","mo":"i","rd":"2026-03-01","lu":"2026-03-01"},{"p":"regolo-ai","id":"qwen3.5-122b","n":"Qwen3.5-122B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.9,"co":3.6,"ctx":262144,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"regolo-ai","id":"gpt-oss-120b","n":"GPT-OSS-120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":1,"co":4.2,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"regolo-ai","id":"qwen3-coder-next","n":"Qwen3-Coder-Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.3,"co":1.2,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"regolo-ai","id":"qwen3.5-9b","n":"Qwen3.5-9B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.15,"co":0.6,"ctx":262144,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"regolo-ai","id":"mistral-small-4-119b","n":"Mistral Small 4 119B","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.75,"co":3,"ctx":256000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"regolo-ai","id":"gpt-oss-20b","n":"GPT-OSS-20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.4,"co":1.8,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"stackit","id":"cortecs/Llama-3.3-70B-Instruct-FP8-Dynamic","n":"Llama 3.3 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.53,"co":0.76,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-12-05","lu":"2024-12-05"},{"p":"stackit","id":"google/gemma-3-27b-it","n":"Gemma 3 27B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.53,"co":0.76,"ctx":37000,"out":4096,"mi":"ti","mo":"t","ow":1,"rd":"2025-05-17","lu":"2025-05-17"},{"p":"stackit","id":"Qwen/Qwen3.6-27B","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.53,"co":0.76,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"stackit","id":"Qwen/Qwen3-VL-Embedding-8B","n":"Qwen3-VL Embedding 8B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.09,"co":0.09,"ctx":32000,"out":4096,"mi":"ti","mo":"t","ow":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"stackit","id":"Qwen/Qwen3-VL-235B-A22B-Instruct-FP8","n":"Qwen3-VL 235B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.76,"co":2.05,"ctx":218000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2024-11-01","lu":"2024-11-01"},{"p":"stackit","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.53,"co":0.76,"ctx":131000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"stackit","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.18,"co":0.29,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"stackit","id":"intfloat/e5-mistral-7b-instruct","n":"E5 Mistral 7B","f":"mistral","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.02,"co":0.02,"ctx":4096,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2023-12-11","lu":"2023-12-11"},{"p":"vercel","id":"xai/grok-imagine-video-1.5","n":"Grok Imagine Video 1.5","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-06-22","lu":"2026-06-22"},{"p":"vercel","id":"xai/grok-4.1-fast-reasoning","n":"Grok 4.1 Fast Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-11-19","lu":"2025-07-09"},{"p":"vercel","id":"xai/grok-4.20-non-reasoning-beta","n":"Grok 4.20 Beta Non-Reasoning","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.4,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-11","lu":"2026-03-13"},{"p":"vercel","id":"xai/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-30","lu":"2026-04-17"},{"p":"vercel","id":"xai/grok-tts","n":"Grok TTS","f":"grok","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2026-03-16","lu":"2026-03-16"},{"p":"vercel","id":"xai/grok-4.1-fast-non-reasoning","n":"Grok 4.1 Fast Non-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","tc":1,"kn":"2024-10","rd":"2025-11-19","lu":"2025-07-09"},{"p":"vercel","id":"xai/grok-voice-think-fast-1.0","n":"Grok Voice Think Fast 1.0","f":"grok","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"ta","mo":"ta","rd":"2026-04-23","lu":"2026-04-23"},{"p":"vercel","id":"xai/grok-imagine-video","n":"Grok Imagine","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-01-28","lu":"2026-01-28"},{"p":"vercel","id":"xai/grok-4.20-multi-agent-beta","n":"Grok 4.20 Multi Agent Beta","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-11","lu":"2026-03-13"},{"p":"vercel","id":"xai/grok-stt","n":"Grok STT","f":"grok","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ctx":0,"out":0,"mi":"a","mo":"t","rd":"2026-03-16","lu":"2026-03-16"},{"p":"vercel","id":"xai/grok-4.20-reasoning","n":"Grok 4.20 Reasoning","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-10","lu":"2026-03-23"},{"p":"vercel","id":"xai/grok-4.20-reasoning-beta","n":"Grok 4.20 Beta Reasoning","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-11","lu":"2026-03-13"},{"p":"vercel","id":"xai/grok-imagine-video-1.5-preview","n":"Grok Imagine Video 1.5 Preview","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-05-30","lu":"2026-05-30"},{"p":"vercel","id":"xai/grok-4.20-non-reasoning","n":"Grok 4.20 Non-Reasoning","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-10","lu":"2026-03-23"},{"p":"vercel","id":"xai/grok-4.20-multi-agent","n":"Grok 4.20 Multi-Agent","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-10","lu":"2026-03-23"},{"p":"vercel","id":"xai/grok-imagine-image","n":"Grok Imagine Image","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"ti","rd":"2026-01-28","lu":"2026-02-19"},{"p":"vercel","id":"xai/grok-build-0.1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-20","lu":"2026-04-16"},{"p":"vercel","id":"moonshotai/kimi-k2","n":"Kimi K2 Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.57,"co":2.3,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2025-07-11","lu":"2025-09-05"},{"p":"vercel","id":"moonshotai/kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":256000,"out":32768,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"vercel","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Thinking Kimi model for slower research passes, planning, and hard technical questions","ci":0.47,"co":2,"cr":0.141,"ctx":216144,"out":216144,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"vercel","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.6,"co":3,"cr":0.1,"ctx":262114,"out":262114,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-26","lu":"2026-01"},{"p":"vercel","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.16,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-20","lu":"2026-04-21"},{"p":"vercel","id":"moonshotai/kimi-k2.7-code-highspeed","n":"Kimi K2.7 Code High Speed","f":"kimi-k2","d":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","ci":1.9,"co":8,"cr":0.38,"ctx":262144,"out":32768,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-06-15","lu":"2026-06-12"},{"p":"vercel","id":"klingai/kling-v3.0-motion-control","n":"Kling v3.0 Motion Control","f":"ling","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-03-04","lu":"2026-03-04"},{"p":"vercel","id":"klingai/kling-v2.6-i2v","n":"Kling v2.6 Image-to-Video","f":"ling","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-03","lu":"2025-12-21"},{"p":"vercel","id":"klingai/kling-v2.5-turbo-t2v","n":"Kling v2.5 Turbo Text-to-Video","f":"ling","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-09-23","lu":"2025-09-23"},{"p":"vercel","id":"klingai/kling-v3.0-i2v","n":"Kling v3.0 Image-to-Video","f":"ling","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-02-05","lu":"2026-02-05"},{"p":"vercel","id":"klingai/kling-v2.5-turbo-i2v","n":"Kling v2.5 Turbo Image-to-Video","f":"ling","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-09-23","lu":"2025-09-23"},{"p":"vercel","id":"klingai/kling-v3.0-t2v","n":"Kling v3.0 Text-to-Video","f":"ling","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-02-05","lu":"2026-02-05"},{"p":"vercel","id":"klingai/kling-v2.6-motion-control","n":"Kling v2.6 Motion Control","f":"ling","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-18","lu":"2025-12-21"},{"p":"vercel","id":"klingai/kling-v2.6-t2v","n":"Kling v2.6 Text-to-Video","f":"ling","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-03","lu":"2025-12-21"},{"p":"vercel","id":"voyage/voyage-4-lite","n":"voyage-4-lite","f":"voyage","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":32000,"out":0,"mi":"t","mo":"t","rd":"2026-01-15","lu":"2026-03-06"},{"p":"vercel","id":"voyage/voyage-law-2","n":"voyage-law-2","f":"voyage","d":"General-purpose chat model for instruction following, writing, and analysis","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-04-15","lu":"2024-03"},{"p":"vercel","id":"voyage/voyage-4","n":"voyage-4","f":"voyage","d":"General-purpose chat model for instruction following, writing, and analysis","ctx":32000,"out":0,"mi":"t","mo":"t","rd":"2026-01-15","lu":"2026-03-06"},{"p":"vercel","id":"voyage/voyage-code-3","n":"voyage-code-3","f":"voyage","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-12-04","lu":"2024-09"},{"p":"vercel","id":"voyage/voyage-4-large","n":"voyage-4-large","f":"voyage","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":32000,"out":0,"mi":"t","mo":"t","rd":"2026-01-15","lu":"2026-03-06"},{"p":"vercel","id":"voyage/rerank-2.5","n":"Voyage Rerank 2.5","f":"voyage","d":"Reranking model for improving retrieval quality in search and recommendation systems","ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2025-08-11","lu":"2025-08-11"},{"p":"vercel","id":"voyage/rerank-2.5-lite","n":"Voyage Rerank 2.5 Lite","f":"voyage","d":"Reranking model for improving retrieval quality in search and recommendation systems","ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2025-08-11","lu":"2025-08-11"},{"p":"vercel","id":"voyage/voyage-code-2","n":"voyage-code-2","f":"voyage","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01"},{"p":"vercel","id":"voyage/voyage-3.5-lite","n":"voyage-3.5-lite","f":"voyage","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2025-05-20","lu":"2025-05-20"},{"p":"vercel","id":"voyage/voyage-3.5","n":"voyage-3.5","f":"voyage","d":"General-purpose chat model for instruction following, writing, and analysis","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2025-05-20","lu":"2025-05-20"},{"p":"vercel","id":"voyage/voyage-3-large","n":"voyage-3-large","f":"voyage","d":"Flagship model for demanding analysis, coding, and production agent workflows","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2025-01-07","lu":"2024-09"},{"p":"vercel","id":"voyage/voyage-finance-2","n":"voyage-finance-2","f":"voyage","d":"General-purpose chat model for instruction following, writing, and analysis","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-06-03","lu":"2024-03"},{"p":"vercel","id":"mistral/mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-07-18","lu":"2024-07-01"},{"p":"vercel","id":"mistral/codestral-embed","n":"Codestral Embed","f":"codestral-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2025-05-28","lu":"2025-05-28"},{"p":"vercel","id":"mistral/mistral-embed","n":"Mistral Embed","f":"mistral-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2023-12-11","lu":"2023-12-11"},{"p":"vercel","id":"mistral/devstral-small","n":"Devstral Small 1.1","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.1,"co":0.3,"ctx":128000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2025-05-21","lu":"2025-05-07"},{"p":"vercel","id":"mistral/mistral-large-3","n":"Mistral Large 3","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0.5,"co":1.5,"ctx":256000,"out":256000,"mi":"ti","mo":"t","kn":"2024-10","rd":"2025-12-02","lu":"2025-12-02"},{"p":"vercel","id":"mistral/mistral-medium-3.5","n":"Mistral Medium Latest","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":1.5,"co":7.5,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-29","lu":"2026-05-21"},{"p":"vercel","id":"mistral/mistral-medium","n":"Mistral Medium 3.1","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":128000,"out":64000,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2025-05-07","lu":"2025-05-07"},{"p":"vercel","id":"mistral/devstral-small-2","n":"Devstral Small 2","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.1,"co":0.3,"ctx":256000,"out":256000,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2025-12-09","lu":"2025-05-07"},{"p":"vercel","id":"mistral/ministral-14b","n":"Ministral 14B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"ctx":256000,"out":256000,"mi":"tip","mo":"t","kn":"2024-10","rd":"2025-12-02","lu":"2025-12-01"},{"p":"vercel","id":"mistral/devstral-2","n":"Devstral 2","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":2,"ctx":256000,"out":256000,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2025-12-09","lu":"2025-12-09"},{"p":"vercel","id":"mistral/mistral-small","n":"Mistral Small (latest)","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"ctx":32000,"out":4000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-06","rd":"2024-09-17","lu":"2026-03-16"},{"p":"vercel","id":"mistral/ministral-8b","n":"Ministral 8B (latest)","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-10-01","lu":"2024-10-04"},{"p":"vercel","id":"mistral/codestral","n":"Codestral (latest)","f":"codestral","d":"Mistral code model for completions, refactors, and developer IDE workflows","ci":0.3,"co":0.9,"ctx":256000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-05-29","lu":"2025-01-04"},{"p":"vercel","id":"mistral/pixtral-12b","n":"Pixtral 12B","f":"pixtral","d":"Mistral vision-language model for image understanding and multimodal chat","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-09","rd":"2024-09-01","lu":"2024-09-01"},{"p":"vercel","id":"mistral/pixtral-large","n":"Pixtral Large (latest)","f":"pixtral","d":"Mistral's larger vision model for document-heavy image understanding and chat","ci":2,"co":6,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2024-11-04"},{"p":"vercel","id":"mistral/ministral-3b","n":"Ministral 3B (latest)","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.04,"co":0.04,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-10-01","lu":"2024-10-04"},{"p":"vercel","id":"mistral/magistral-small","n":"Magistral Small","f":"magistral-small","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":0.5,"co":1.5,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-03-17","lu":"2025-03-17"},{"p":"vercel","id":"mistral/magistral-medium","n":"Magistral Medium (latest)","f":"magistral-medium","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":2,"co":5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-03-17","lu":"2025-03-20"},{"p":"vercel","id":"google/gemini-embedding-2","n":"Gemini Embedding 2","f":"gemini-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":0,"out":0,"mi":"t","mo":"t","rd":"2026-03-10","lu":"2026-03-23"},{"p":"vercel","id":"google/text-multilingual-embedding-002","n":"Text Multilingual Embedding 002","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-03-01","lu":"2024-03"},{"p":"vercel","id":"google/gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.03,"ctx":1000000,"out":65000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"vercel","id":"google/gemini-3.1-flash-image","n":"Gemini 3.1 Flash Image (Nano Banana 2)","f":"gemini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":3,"cr":0.05,"ctx":131072,"out":32768,"mi":"ti","mo":"ti","rs":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"vercel","id":"google/gemini-3-flash","n":"Gemini 3 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"ctx":1000000,"out":65000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2025-12-17","lu":"2025-12-17"},{"p":"vercel","id":"google/veo-3.1-generate-001","n":"Veo 3.1","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-10-15","lu":"2026-06-08"},{"p":"vercel","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"vercel","id":"google/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.14,"co":0.4,"ctx":262144,"out":131072,"mi":"tip","mo":"t","tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"vercel","id":"google/veo-3.0-generate-001","n":"Veo 3.0","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-05-20","lu":"2026-06-08"},{"p":"vercel","id":"google/veo-3.0-fast-generate-001","n":"Veo 3.0 Fast Generate","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-07-31","lu":"2026-06-08"},{"p":"vercel","id":"google/text-embedding-005","n":"Text Embedding 005","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-08-01","lu":"2024-08"},{"p":"vercel","id":"google/gemini-embedding-001","n":"Gemini Embedding 001","f":"gemini-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","kn":"2025-05","rd":"2025-05-20","lu":"2025-05-20"},{"p":"vercel","id":"google/gemini-3.1-flash-lite-image","n":"Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite)","f":"gemini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.25,"co":1.5,"cr":0.03,"ctx":65536,"out":4096,"mi":"ti","mo":"ti","rs":1,"rd":"2026-06-30","lu":"2026-06-30"},{"p":"vercel","id":"google/gemini-2.5-flash-image","n":"Nano Banana (Gemini 2.5 Flash Image)","f":"gemini-flash","d":"Nano Banana image model for fast generation, edits, and character-consistent assets","ci":0.3,"co":2.5,"cr":0.03,"ctx":32768,"out":65536,"mi":"t","mo":"ti","kn":"2025-01","rd":"2025-08-26","lu":"2025-08-26"},{"p":"vercel","id":"google/imagen-4.0-fast-generate-001","n":"Imagen 4 Fast","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","rd":"2025-06-01","lu":"2025-06"},{"p":"vercel","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"vercel","id":"google/gemini-3.1-flash-image-preview","n":"Gemini 3.1 Flash Image Preview (Nano Banana 2)","f":"gemini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":3,"cr":0.05,"ctx":131072,"out":32768,"mi":"ti","mo":"ti","rs":1,"kn":"2025-01","rd":"2026-02-26","lu":"2026-02-26"},{"p":"vercel","id":"google/imagen-4.0-ultra-generate-001","n":"Imagen 4 Ultra","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","rd":"2025-05-24","lu":"2025-05-24"},{"p":"vercel","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2026-02-19"},{"p":"vercel","id":"google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.15,"co":0.6,"cr":0.015,"ctx":262144,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"vercel","id":"google/gemini-3-pro-preview","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":2,"co":12,"cr":0.2,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"vercel","id":"google/veo-3.1-fast-generate-001","n":"Veo 3.1 Fast Generate","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-10-15","lu":"2026-06-08"},{"p":"vercel","id":"google/gemini-3-pro-image","n":"Nano Banana Pro (Gemini 3 Pro Image)","f":"gemini-pro","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":12,"cr":0.2,"ctx":65536,"out":32768,"mi":"t","mo":"ti","kn":"2025-03","rd":"2025-09-01","lu":"2025-09"},{"p":"vercel","id":"google/gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.03,"ctx":1000000,"out":65000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"vercel","id":"google/imagen-4.0-generate-001","n":"Imagen 4","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","rd":"2025-05-20","lu":"2025-05-22"},{"p":"vercel","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"vercel","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"vercel","id":"prodia/flux-fast-schnell","n":"Flux Schnell","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2024-08-02","lu":"2026-06-08"},{"p":"vercel","id":"openai/gpt-oss-safeguard-20b","n":"gpt-oss-safeguard-20b","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.075,"co":0.3,"cr":0.037,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-10-29","lu":"2024-12-01"},{"p":"vercel","id":"openai/gpt-3.5-turbo-instruct","n":"GPT-3.5 Turbo Instruct","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":8192,"out":4096,"mi":"t","mo":"t","kn":"2021-09","rd":"2023-09-18","lu":"2023-03-01"},{"p":"vercel","id":"openai/gpt-5.2-chat","n":"GPT-5.2 Chat","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2024-10","rd":"2025-12-11","lu":"2025-08-07"},{"p":"vercel","id":"openai/text-embedding-3-large","n":"text-embedding-3-large","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-01-25","lu":"2024-01-25"},{"p":"vercel","id":"openai/gpt-5.2-pro","n":"GPT 5.2 ","f":"gpt","d":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-11","lu":"2025-12-11"},{"p":"vercel","id":"openai/gpt-4o-mini-search-preview","n":"GPT 4o Mini Search Preview","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"t","mo":"t","kn":"2023-09","rd":"2025-03-12","lu":"2025-01"},{"p":"vercel","id":"openai/gpt-5-chat","n":"GPT-5 Chat","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"tip","mo":"ti","rs":1,"tc":1,"kn":"2024-10","rd":"2025-08-07","lu":"2025-08-07"},{"p":"vercel","id":"openai/gpt-3.5-turbo","n":"GPT-3.5 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"ctx":16385,"out":4096,"mi":"t","mo":"t","kn":"2021-09","rd":"2023-03-01","lu":"2023-11-06"},{"p":"vercel","id":"openai/gpt-5-pro","n":"GPT-5 pro","f":"gpt","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-10-06","lu":"2025-10-06"},{"p":"vercel","id":"openai/o3-pro","n":"o3 Pro","f":"o-pro","d":"High-effort o3 tier for difficult technical reasoning and careful answers","ci":20,"co":80,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-06-10","lu":"2025-06-10"},{"p":"vercel","id":"openai/gpt-4o-transcribe","n":"GPT-4o Transcribe","f":"gpt","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":2.5,"co":10,"ctx":0,"out":0,"mi":"a","mo":"t","rd":"2024-03-13","lu":"2024-03-13"},{"p":"vercel","id":"openai/gpt-5.4-nano","n":"GPT 5.4 Nano","f":"gpt","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"vercel","id":"openai/gpt-5.3-chat","n":"GPT-5.3 Chat","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-03","lu":"2026-03-06"},{"p":"vercel","id":"openai/gpt-5.1-thinking","n":"GPT 5.1 Thinking","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"tip","mo":"ti","rs":1,"tc":1,"kn":"2024-10","rd":"2025-11-12","lu":"2025-08-07"},{"p":"vercel","id":"openai/gpt-5.1-codex","n":"GPT-5.1-Codex","f":"gpt","d":"Codex GPT for repository edits, code review, and practical software agents","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-11-12","lu":"2025-11-13"},{"p":"vercel","id":"openai/gpt-5.1-codex-max","n":"GPT 5.1 Codex Max","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-11-19","lu":"2025-11-13"},{"p":"vercel","id":"openai/gpt-4o-mini-transcribe","n":"GPT-4o mini Transcribe","f":"o-mini","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":1.25,"co":5,"ctx":0,"out":0,"mi":"a","mo":"t","rd":"2024-03-13","lu":"2024-03-13"},{"p":"vercel","id":"openai/text-embedding-ada-002","n":"text-embedding-ada-002","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2022-12-15","lu":"2022-12-15"},{"p":"vercel","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-11","lu":"2025-12-11"},{"p":"vercel","id":"openai/gpt-5.3-codex","n":"GPT 5.3 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"vercel","id":"openai/text-embedding-3-small","n":"text-embedding-3-small","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-01-25","lu":"2024-01-25"},{"p":"vercel","id":"openai/gpt-realtime-1.5","n":"GPT-Realtime-1.5","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":4,"co":16,"cr":0.4,"ctx":0,"out":0,"mi":"ta","mo":"ta","rd":"2026-02-23","lu":"2026-02-23"},{"p":"vercel","id":"openai/gpt-5.1-codex-mini","n":"GPT-5.1 Codex mini","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-11-12","lu":"2025-11-13"},{"p":"vercel","id":"openai/gpt-image-1.5","n":"GPT Image 1.5","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":32,"cr":1.25,"ctx":0,"out":0,"mi":"t","mo":"i","rd":"2025-12-16","lu":"2025-11-25"},{"p":"vercel","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.1,"co":0.5,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2025-08-05","lu":"2025-08-05"},{"p":"vercel","id":"openai/gpt-5.4","n":"GPT 5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"vercel","id":"openai/gpt-5.4-mini","n":"GPT 5.4 Mini","f":"gpt","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"vercel","id":"openai/gpt-realtime-mini","n":"GPT-Realtime mini","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0.6,"co":2.4,"cr":0.06,"ctx":0,"out":0,"mi":"ta","mo":"ta","rd":"2025-10-10","lu":"2025-10-10"},{"p":"vercel","id":"openai/tts-1","n":"TTS-1","f":"o","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2023-11-06","lu":"2023-11-06"},{"p":"vercel","id":"openai/whisper-1","n":"Whisper","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ctx":0,"out":0,"mi":"a","mo":"t","rd":"2022-09-21","lu":"2022-09-21"},{"p":"vercel","id":"openai/tts-1-hd","n":"TTS-1 HD","f":"o","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2023-11-06","lu":"2023-11-06"},{"p":"vercel","id":"openai/o3-deep-research","n":"o3-deep-research","f":"o","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":10,"co":40,"cr":2.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-06-26","lu":"2024-06-26"},{"p":"vercel","id":"openai/gpt-image-1","n":"GPT Image 1","f":"gpt-image","d":"OpenAI image model for production generation, edits, and brand-safe visual workflows","ci":5,"co":40,"cr":1.25,"ctx":0,"out":0,"mi":"t","mo":"i","rd":"2025-03-25","lu":"2025-04-24"},{"p":"vercel","id":"openai/gpt-realtime-2","n":"gpt-realtime-2","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":4,"co":24,"cr":0.4,"ctx":0,"out":0,"mi":"ta","mo":"ta","rd":"2026-05-07","lu":"2026-05-07"},{"p":"vercel","id":"openai/gpt-image-1-mini","n":"GPT Image 1 Mini","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":8,"cr":0.2,"ctx":0,"out":0,"mi":"t","mo":"i","rd":"2025-10-06","lu":"2025-10-06"},{"p":"vercel","id":"openai/gpt-5.4-pro","n":"GPT 5.4 Pro","f":"gpt","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"vercel","id":"openai/gpt-5.5-pro","n":"GPT 5.5 Pro","f":"gpt","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-24","lu":"2026-04-23"},{"p":"vercel","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.2,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2025-08-05","lu":"2025-08-05"},{"p":"vercel","id":"openai/gpt-5.2-codex","n":"GPT-5.2-Codex","f":"gpt-codex","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-18","lu":"2025-12-11"},{"p":"vercel","id":"openai/gpt-5.1-instant","n":"GPT-5.1 Instant","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2024-10","rd":"2025-11-12","lu":"2025-08-07"},{"p":"vercel","id":"openai/gpt-image-2","n":"GPT Image 2","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":30,"cr":1.25,"ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-04-21","lu":"2026-04-21"},{"p":"vercel","id":"openai/gpt-5.5","n":"GPT 5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-24","lu":"2026-04-23"},{"p":"vercel","id":"openai/gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"vercel","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"vercel","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"vercel","id":"openai/gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"vercel","id":"openai/gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"vercel","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"vercel","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"vercel","id":"openai/o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"vercel","id":"openai/gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"vercel","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"vercel","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"vercel","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"vercel","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"vercel","id":"openai/o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"vercel","id":"zai/glm-4.7","n":"GLM 4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.6,"co":2.2,"cr":0.12,"ctx":200000,"out":120000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-22","lu":"2025-12-22"},{"p":"vercel","id":"zai/glm-4.5v","n":"GLM 4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"cr":0.11,"ctx":66000,"out":16000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-11","lu":"2025-08-11"},{"p":"vercel","id":"zai/glm-4.5","n":"GLM 4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"ctx":128000,"out":96000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-28","lu":"2025-07-28"},{"p":"vercel","id":"zai/glm-4.7-flashx","n":"GLM 4.7 FlashX","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.06,"co":0.4,"cr":0.01,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-19","lu":"2026-01-19"},{"p":"vercel","id":"zai/glm-5.1","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.3,"co":4.3,"cr":0.26,"ctx":202000,"out":202000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"vercel","id":"zai/glm-4.6","n":"GLM 4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.6,"co":2.2,"cr":0.11,"ctx":200000,"out":96000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"vercel","id":"zai/glm-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1040000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-16","lu":"2026-06-13"},{"p":"vercel","id":"zai/glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"cr":0.05,"ctx":128000,"out":24000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-09-30","lu":"2025-12-08"},{"p":"vercel","id":"zai/glm-5.2-fast","n":"GLM 5.2 Fast","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":3,"co":10.25,"cr":0.5,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-06-16","lu":"2026-06-16"},{"p":"vercel","id":"zai/glm-5v-turbo","n":"GLM 5V Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":1.2,"co":4,"cr":0.24,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"vercel","id":"zai/glm-4.6v-flash","n":"GLM-4.6V-Flash","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ctx":128000,"out":24000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-09-30","lu":"2025-09-30"},{"p":"vercel","id":"zai/glm-4.5-air","n":"GLM 4.5 Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.2,"co":1.1,"cr":0.03,"ctx":128000,"out":96000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"vercel","id":"zai/glm-4.7-flash","n":"GLM 4.7 Flash","f":"glm","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0.07,"co":0.4,"ctx":200000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"vercel","id":"zai/glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.95,"co":3.15,"cr":0.2,"ctx":202800,"out":131100,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"vercel","id":"zai/glm-5-turbo","n":"GLM 5 Turbo","f":"glm","d":"Faster GLM-5 lane for coding agents that need lower latency","ci":1.2,"co":4,"cr":0.24,"ctx":202800,"out":131100,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-16"},{"p":"vercel","id":"bytedance/seedream-5.0-lite","n":"Seedream 5.0 Lite","f":"seed","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-02-13","lu":"2026-01-28"},{"p":"vercel","id":"bytedance/seedance-2.0-fast","n":"Seedance 2.0 Fast","f":"seed","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"ti","mo":"v","rd":"2026-04-14","lu":"2026-04-14"},{"p":"vercel","id":"bytedance/seedance-v1.0-pro","n":"Seedance v1.0 Pro","f":"seed","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-06-11","lu":"2025-06-11"},{"p":"vercel","id":"bytedance/seed-1.6","n":"Seed 1.6","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"cr":0.05,"ctx":256000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-09-01","lu":"2025-09"},{"p":"vercel","id":"bytedance/seedance-2.0","n":"Seedance 2.0","f":"seed","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"ti","mo":"v","rd":"2026-04-14","lu":"2026-04-14"},{"p":"vercel","id":"bytedance/seedance-v1.5-pro","n":"Seedance v1.5 Pro","f":"seed","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-16","lu":"2025-12-16"},{"p":"vercel","id":"bytedance/seedance-v1.0-pro-fast","n":"Seedance v1.0 Pro Fast","f":"seed","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-10-24","lu":"2025-10-31"},{"p":"vercel","id":"bytedance/seedream-4.0","n":"Seedream 4.0","f":"seed","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2025-09-09","lu":"2025-08-28"},{"p":"vercel","id":"bytedance/seedream-4.5","n":"Seedream 4.5","f":"seed","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2025-12-03","lu":"2025-11-28"},{"p":"vercel","id":"bytedance/seed-1.8","n":"Seed 1.8","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"cr":0.05,"ctx":256000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-09-01","lu":"2025-10"},{"p":"vercel","id":"morph/morph-v3-large","n":"Morph v3 Large","f":"morph","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.9,"co":1.9,"ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2024-08-15","lu":"2024-08-15"},{"p":"vercel","id":"morph/morph-v3-fast","n":"Morph v3 Fast","f":"morph","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.8,"co":1.2,"ctx":16000,"out":16000,"mi":"t","mo":"t","rd":"2024-08-15","lu":"2024-08-15"},{"p":"vercel","id":"sakana/fugu-ultra","n":"Fugu Ultra","f":"aura","d":"Quality-first multi-agent model for hard research, analysis, and competitions","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-21","lu":"2026-06-21"},{"p":"vercel","id":"nvidia/nemotron-3-ultra-550b-a55b","n":"Nemotron 3 Ultra","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.6,"co":2.4,"cr":0.12,"ctx":1000000,"out":65000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"vercel","id":"nvidia/nemotron-3-nano-30b-a3b","n":"Nemotron 3 Nano 30B A3B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.05,"co":0.24,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"kn":"2024-10","rd":"2025-12-15","lu":"2025-12-15"},{"p":"vercel","id":"nvidia/nemotron-nano-9b-v2","n":"Nvidia Nemotron Nano 9B V2","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0.06,"co":0.23,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-08-18","lu":"2025-08-18"},{"p":"vercel","id":"nvidia/nemotron-3-super-120b-a12b","n":"NVIDIA Nemotron 3 Super 120B A12B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.15,"co":0.65,"ctx":256000,"out":32000,"mi":"t","mo":"t","rs":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"vercel","id":"nvidia/nemotron-nano-12b-v2-vl","n":"Nvidia Nemotron Nano 12B V2 VL","f":"nemotron","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0.2,"co":0.6,"ctx":131072,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-10-28","lu":"2025-10-28"},{"p":"vercel","id":"xiaomi/mimo-v2.5","n":"MiMo M2.5","f":"mimo-v2.5","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1050000,"out":131100,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"vercel","id":"xiaomi/mimo-v2-flash","n":"MiMo V2 Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"cr":0.01,"ctx":262144,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-16","lu":"2026-02-04"},{"p":"vercel","id":"xiaomi/mimo-v2-pro","n":"MiMo V2 Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":1,"co":3,"cr":0.2,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"vercel","id":"xiaomi/mimo-v2.5-pro","n":"MiMo V2.5 Pro","f":"mimo-v2.5-pro","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1050000,"out":131000,"mi":"tp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"vercel","id":"quiverai/arrow-1.1","n":"Arrow 1.1","f":"o","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":131072,"out":131072,"mi":"t","mo":"i","rd":"2026-04-16","lu":"2026-04-16"},{"p":"vercel","id":"inception/mercury-coder-small","n":"Mercury Coder Small Beta","f":"mercury","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.25,"co":1,"ctx":32000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-02-26","lu":"2025-02-26"},{"p":"vercel","id":"inception/mercury-2","n":"Mercury 2","f":"mercury","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.25,"co":0.75,"cr":0.024999999999999998,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-24","lu":"2026-03-06"},{"p":"vercel","id":"anthropic/claude-3.5-haiku","n":"Claude 3.5 Haiku","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-11-04","lu":"2024-10-22"},{"p":"vercel","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"vercel","id":"anthropic/claude-haiku-4.5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"vercel","id":"anthropic/claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-29","lu":"2026-06-30"},{"p":"vercel","id":"anthropic/claude-opus-4.7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"vercel","id":"anthropic/claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"vercel","id":"anthropic/claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-07-01","lu":"2026-06-09"},{"p":"vercel","id":"anthropic/claude-opus-4.5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"vercel","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"vercel","id":"anthropic/claude-3-haiku","n":"Claude Haiku 3","f":"claude-haiku","d":"Legacy model retained for compatibility with older integrations","ci":0.25,"co":1.25,"cr":0.03,"cw":0.3,"ctx":200000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-03-13","lu":"2024-03-13"},{"p":"vercel","id":"anthropic/claude-opus-4","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"vercel","id":"anthropic/claude-opus-4.6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"vercel","id":"anthropic/claude-opus-4.1","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"vercel","id":"anthropic/claude-sonnet-4.5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"vercel","id":"cohere/rerank-v3.5","n":"Cohere Rerank 3.5","f":"o","d":"Reranking model for improving retrieval quality in search and recommendation systems","ctx":4096,"out":4096,"mi":"t","mo":"t","rd":"2024-12-02","lu":"2024-12-02"},{"p":"vercel","id":"cohere/command-a","n":"Command A","f":"command","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8000,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2025-03-13","lu":"2025-03-13"},{"p":"vercel","id":"cohere/rerank-v4-fast","n":"Cohere Rerank 4 Fast","f":"o","d":"Reranking model for improving retrieval quality in search and recommendation systems","ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2025-12-11","lu":"2025-12-11"},{"p":"vercel","id":"cohere/embed-v4.0","n":"Embed v4.0","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":128000,"out":1536,"mi":"t","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"vercel","id":"cohere/rerank-v4-pro","n":"Cohere Rerank 4 Pro","f":"o","d":"Reranking model for improving retrieval quality in search and recommendation systems","ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2025-12-11","lu":"2025-12-11"},{"p":"vercel","id":"stepfun/step-3.7-flash","n":"Step 3.7 Flash","f":"step","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0.2,"co":1.15,"cr":0.04,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-01-01","rd":"2026-05-28","lu":"2026-05-29"},{"p":"vercel","id":"stepfun/step-3.5-flash","n":"StepFun 3.5 Flash","f":"step","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.09,"co":0.3,"cr":0.02,"ctx":262114,"out":262114,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-02-13"},{"p":"vercel","id":"interfaze/interfaze-beta","n":"Interfaze Beta","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":1.5,"co":3.5,"ctx":1000000,"out":32000,"mi":"tip","mo":"t","rs":1,"rd":"2025-10-07","lu":"2026-04-29"},{"p":"vercel","id":"bfl/flux-kontext-max","n":"FLUX.1 Kontext Max","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2025-05-29","lu":"2025-06"},{"p":"vercel","id":"bfl/flux-2-flex","n":"FLUX.2 [flex]","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2025-11-25","lu":"2026-06-08"},{"p":"vercel","id":"bfl/flux-pro-1.1-ultra","n":"FLUX1.1 [pro] Ultra","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2024-11-01","lu":"2024-11"},{"p":"vercel","id":"bfl/flux-2-max","n":"FLUX.2 [max]","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":67300,"out":67300,"mi":"t","mo":"i","rd":"2025-12-16","lu":"2026-06-08"},{"p":"vercel","id":"bfl/flux-pro-1.1","n":"FLUX1.1 [pro]","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2024-10-02","lu":"2024-10"},{"p":"vercel","id":"bfl/flux-pro-1.0-fill","n":"FLUX.1 Fill [pro]","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2024-10-01","lu":"2024-10"},{"p":"vercel","id":"bfl/flux-2-klein-4b","n":"FLUX.2 [klein] 4B","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-01-15","lu":"2026-06-08"},{"p":"vercel","id":"bfl/flux-2-klein-9b","n":"FLUX.2 [klein] 9B","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-01-15","lu":"2026-06-08"},{"p":"vercel","id":"bfl/flux-kontext-pro","n":"FLUX.1 Kontext Pro","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2025-05-29","lu":"2025-06"},{"p":"vercel","id":"bfl/flux-2-pro","n":"FLUX.2 [pro]","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":67300,"out":67300,"mi":"t","mo":"i","rd":"2025-11-25","lu":"2026-06-08"},{"p":"vercel","id":"recraft/recraft-v4.1-pro","n":"Recraft V4.1 Pro","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-05-14","lu":"2026-05-14"},{"p":"vercel","id":"recraft/recraft-v4.1","n":"Recraft V4.1","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-05-14","lu":"2026-05-14"},{"p":"vercel","id":"recraft/recraft-v4","n":"Recraft V4","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-02-17","lu":"2026-02-17"},{"p":"vercel","id":"recraft/recraft-v4-pro","n":"Recraft V4 Pro","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-02-17","lu":"2026-02-17"},{"p":"vercel","id":"recraft/recraft-v2","n":"Recraft V2","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2024-03-13","lu":"2024-03"},{"p":"vercel","id":"recraft/recraft-v3","n":"Recraft V3","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":512,"out":0,"mi":"t","mo":"i","rd":"2024-10-30","lu":"2024-10"},{"p":"vercel","id":"recraft/recraft-v4.1-utility-pro","n":"Recraft V4.1 Utility Pro","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-05-14","lu":"2026-05-14"},{"p":"vercel","id":"recraft/recraft-v4.1-utility","n":"Recraft V4.1 Utility","f":"recraft","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","rd":"2026-05-14","lu":"2026-05-14"},{"p":"vercel","id":"arcee-ai/trinity-large-preview","n":"Trinity Large Preview","f":"trinity","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":1,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2026-01-27","lu":"2025-01"},{"p":"vercel","id":"arcee-ai/trinity-large-thinking","n":"Trinity Large Thinking","f":"trinity","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":0.8999999999999999,"ctx":262100,"out":80000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-03"},{"p":"vercel","id":"arcee-ai/trinity-mini","n":"Trinity Mini","f":"trinity","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.045,"co":0.15,"ctx":131072,"out":131072,"mi":"t","mo":"t","kn":"2024-10","rd":"2025-12-01","lu":"2025-12"},{"p":"vercel","id":"perplexity/sonar-reasoning-pro","n":"Sonar Reasoning Pro","f":"sonar-reasoning","d":"Web-grounded reasoning model for multi-step research and cited answers","ctx":127000,"out":8000,"mi":"t","mo":"t","rs":1,"kn":"2025-09","rd":"2025-02-19","lu":"2025-02-19"},{"p":"vercel","id":"perplexity/sonar","n":"Sonar","f":"sonar","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ctx":127000,"out":8000,"mi":"ti","mo":"t","tc":1,"kn":"2025-02","rd":"2025-02-19","lu":"2025-02-19"},{"p":"vercel","id":"perplexity/sonar-pro","n":"Sonar Pro","f":"sonar-pro","d":"Advanced Sonar search model for deeper research and cited synthesis","ctx":200000,"out":8000,"mi":"ti","mo":"t","tc":1,"kn":"2025-09","rd":"2025-02-19","lu":"2025-02-19"},{"p":"vercel","id":"amazon/titan-embed-text-v2","n":"Titan Text Embeddings V2","f":"titan-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2024-04-30","lu":"2024-04"},{"p":"vercel","id":"amazon/nova-2-lite","n":"Nova 2 Lite","f":"nova","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.3,"co":2.5,"cr":0.075,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"kn":"2024-10","rd":"2025-12-02","lu":"2024-12-01"},{"p":"vercel","id":"amazon/nova-lite","n":"Nova Lite","f":"nova-lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.06,"co":0.24,"cr":0.015,"ctx":300000,"out":8192,"mi":"tiv","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-03","lu":"2024-12-03"},{"p":"vercel","id":"amazon/nova-micro","n":"Nova Micro","f":"nova-micro","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.035,"co":0.14,"cr":0.00875,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-03","lu":"2024-12-03"},{"p":"vercel","id":"amazon/nova-pro","n":"Nova Pro","f":"nova-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.8,"co":3.2,"cr":0.2,"ctx":300000,"out":8192,"mi":"tiv","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-03","lu":"2024-12-03"},{"p":"vercel","id":"alibaba/qwen3-vl-thinking","n":"Qwen3 VL Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":4,"ctx":131072,"out":32768,"mi":"tip","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2025-09-23","lu":"2025-09-24"},{"p":"vercel","id":"alibaba/qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Hosted Qwen coder for software agents, repo edits, and long-context code","ci":1,"co":5,"cr":0.2,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"vercel","id":"alibaba/wan-v2.6-r2v","n":"Wan v2.6 Reference-to-Video","f":"o","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-16","lu":"2025-12-16"},{"p":"vercel","id":"alibaba/qwen3-embedding-0.6b","n":"Qwen3 Embedding 0.6B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"vercel","id":"alibaba/qwen3-max-preview","n":"Qwen3 Max Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.2,"co":6,"cr":0.24,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-05","lu":"2025-09-23"},{"p":"vercel","id":"alibaba/qwen3-embedding-8b","n":"Qwen3 Embedding 8B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-06-05","lu":"2025-06-05"},{"p":"vercel","id":"alibaba/qwen3-next-80b-a3b-instruct","n":"Qwen3 Next 80B A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":1.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-11","lu":"2025-09"},{"p":"vercel","id":"alibaba/qwen3.7-plus","n":"Qwen 3.7 Plus","f":"qwen3.7-plus","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.4,"co":1.6,"cr":0.08,"cw":0.5,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"vercel","id":"alibaba/wan-v2.7-r2v","n":"Wan v2.7 Reference-to-Video","f":"o","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-04-07","lu":"2026-04-07"},{"p":"vercel","id":"alibaba/qwen3-vl-instruct","n":"Qwen3 VL Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":1.6,"ctx":131072,"out":129024,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-24"},{"p":"vercel","id":"alibaba/qwen3.7-max","n":"Qwen 3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":1.25,"co":3.75,"cr":0.25,"cw":1.5625,"ctx":991000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"vercel","id":"alibaba/wan-v2.5-t2v-preview","n":"Wan v2.5 Text-to-Video Preview","f":"o","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-09-24","lu":"2025-09-24"},{"p":"vercel","id":"alibaba/qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","ci":1.2,"co":6,"cr":0.24,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"vercel","id":"alibaba/qwen3-next-80b-a3b-thinking","n":"Qwen3 Next 80B A3B Thinking","f":"qwen","d":"Efficient Qwen thinking model for local reasoning, math, and coding agents","ci":0.15,"co":1.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2025-09-11","lu":"2025-09"},{"p":"vercel","id":"alibaba/qwen3-embedding-4b","n":"Qwen3 Embedding 4B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-06-05","lu":"2025-06-05"},{"p":"vercel","id":"alibaba/qwen3.5-flash","n":"Qwen 3.5 Flash","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1,"co":0.4,"cr":0.001,"cw":0.125,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"vercel","id":"alibaba/qwen3-coder","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":1.5,"co":7.5,"cr":0.3,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-07-22","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen-3-235b","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.22,"co":0.88,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-04-28","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen3.5-plus","n":"Qwen 3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":2.4,"cr":0.04,"cw":0.5,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"vercel","id":"alibaba/wan-v2.6-t2v","n":"Wan v2.6 Text-to-Video","f":"o","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-16","lu":"2025-12-16"},{"p":"vercel","id":"alibaba/qwen3-max-thinking","n":"Qwen 3 Max Thinking","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":1.2,"co":6,"cr":0.24,"ctx":256000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-23","lu":"2025-01"},{"p":"vercel","id":"alibaba/wan-v2.6-i2v-flash","n":"Wan v2.6 Image-to-Video Flash","f":"o","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-16","lu":"2025-12-16"},{"p":"vercel","id":"alibaba/wan-v2.6-r2v-flash","n":"Wan v2.6 Reference-to-Video Flash","f":"o","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-16","lu":"2025-12-16"},{"p":"vercel","id":"alibaba/qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.5,"co":1.2,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-07-22","lu":"2026-02-19"},{"p":"vercel","id":"alibaba/qwen3.6-27b","n":"Qwen 3.6 27B","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":256000,"out":256000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"vercel","id":"alibaba/wan-v2.6-i2v","n":"Wan v2.6 Image-to-Video","f":"o","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2025-12-16","lu":"2025-12-16"},{"p":"vercel","id":"alibaba/wan-v2.7-t2v","n":"Wan v2.7 Text-to-Video","f":"o","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"t","mo":"v","rd":"2026-04-07","lu":"2026-04-07"},{"p":"vercel","id":"alibaba/qwen-3-30b","n":"Qwen3-30B-A3B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.12,"co":0.5,"ctx":40960,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-04-28","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen3-235b-a22b-thinking","n":"Qwen3 235B A22B Thinking 2507","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":4,"ctx":131072,"out":32768,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen3-vl-235b-a22b-instruct","n":"Qwen3 VL 235B A22B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":1.6,"ctx":131072,"out":129024,"mi":"ti","mo":"t","rd":"2025-09-23","lu":"2026-05-01"},{"p":"vercel","id":"alibaba/qwen-3-14b","n":"Qwen3-14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.12,"co":0.24,"ctx":40960,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-04-28","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen-3-32b","n":"Qwen 3.32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.16,"co":0.64,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-04-28","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen-3.6-max-preview","n":"Qwen 3.6 Max Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.3,"co":7.8,"cr":0.26,"cw":1.625,"ctx":240000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-20","lu":"2026-04-24"},{"p":"vercel","id":"alibaba/qwen3-coder-30b-a3b","n":"Qwen 3 Coder 30B A3B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.15,"co":0.6,"ctx":262144,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-07-31","lu":"2025-04"},{"p":"vercel","id":"alibaba/qwen3.6-plus","n":"Qwen 3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.5,"co":3,"cr":0.1,"cw":0.625,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"vercel","id":"meituan/longcat-flash-thinking-2601","n":"LongCat Flash Thinking 2601","f":"longcat","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2026-01-15","lu":"2026-03-13"},{"p":"vercel","id":"meituan/longcat-flash-chat","n":"LongCat Flash Chat","f":"longcat","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":128000,"out":100000,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2025-09-01","lu":"2025-08-30"},{"p":"vercel","id":"meta/llama-3.2-1b","n":"Llama 3.2 1B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.1,"co":0.1,"ctx":128000,"out":8192,"mi":"t","mo":"t","kn":"2023-12","rd":"2024-09-25","lu":"2024-09-18"},{"p":"vercel","id":"meta/llama-3.2-11b","n":"Llama 3.2 11B Vision Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.16,"co":0.16,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"vercel","id":"meta/llama-3.1-8b","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.22,"co":0.22,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"vercel","id":"meta/llama-3.2-90b","n":"Llama 3.2 90B Vision Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.72,"co":0.72,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"vercel","id":"meta/llama-3.1-70b","n":"Llama 3.1 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.72,"co":0.72,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"vercel","id":"meta/llama-3.2-3b","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.15,"co":0.15,"ctx":128000,"out":8192,"mi":"t","mo":"t","kn":"2023-12","rd":"2024-09-25","lu":"2024-09-18"},{"p":"vercel","id":"meta/llama-4-scout","n":"Llama-4-Scout-17B-16E-Instruct-FP8","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"vercel","id":"meta/llama-3.3-70b","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"vercel","id":"meta/llama-4-maverick","n":"Llama-4-Maverick-17B-128E-Instruct-FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"vercel","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-23","lu":"2026-04-24"},{"p":"vercel","id":"deepseek/deepseek-v3.1-terminus","n":"DeepSeek V3.1 Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"cr":0.135,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-09-22","lu":"2025-09-22"},{"p":"vercel","id":"deepseek/deepseek-v3","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1.12,"cr":0.135,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-12-26","lu":"2024-12-26"},{"p":"vercel","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-23","lu":"2026-04-24"},{"p":"vercel","id":"deepseek/deepseek-v3.2-thinking","n":"DeepSeek V3.2 Thinking","f":"deepseek-thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.62,"co":1.85,"ctx":128000,"out":8000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"vercel","id":"deepseek/deepseek-v3.1","n":"DeepSeek-V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.6,"co":1.7,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2025-08-21","lu":"2025-08-21"},{"p":"vercel","id":"deepseek/deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":0.42,"cr":0.028,"ctx":128000,"out":8000,"mi":"t","mo":"t","kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"vercel","id":"deepseek/deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":1.35,"co":5.4,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"vercel","id":"minimax/minimax-m2.7-highspeed","n":"MiniMax M2.7 High Speed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131100,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"vercel","id":"minimax/minimax-m2.5","n":"MiniMax M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"vercel","id":"minimax/minimax-m2.1","n":"MiniMax M2.1","f":"minimax","d":"Earlier MiniMax agent model for practical coding and productivity tasks","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-23","lu":"2025-12-23"},{"p":"vercel","id":"minimax/minimax-m3","n":"MiniMax M3","f":"minimax-m3","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.3,"co":1.2,"cr":0.06,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-05-31","lu":"2026-06-01"},{"p":"vercel","id":"minimax/minimax-m2","n":"MiniMax M2","f":"minimax","d":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":205000,"out":205000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2025-10-27","lu":"2025-10-27"},{"p":"vercel","id":"minimax/minimax-m2.7","n":"Minimax M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"vercel","id":"minimax/minimax-m2.5-highspeed","n":"MiniMax M2.5 High Speed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.03,"cw":0.375,"ctx":204800,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-13"},{"p":"vercel","id":"minimax/minimax-m2.1-lightning","n":"MiniMax M2.1 Lightning","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.3,"co":2.4,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-12-23","lu":"2025-10-27"},{"p":"vercel","id":"kwaipilot/kat-coder-pro-v1","n":"KAT-Coder-Pro V1","f":"kat-coder","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":256000,"out":32000,"mi":"t","mo":"t","rs":1,"kn":"2024-10","rd":"2025-11-09","lu":"2025-10-24"},{"p":"vercel","id":"kwaipilot/kat-coder-pro-v2","n":"Kat Coder Pro V2","f":"kat-coder","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-27","lu":"2026-03-30"},{"p":"submodel","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen3 235B A22B Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.2,"co":0.6,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"submodel","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.2,"co":0.8,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"submodel","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.2,"co":0.3,"ctx":262144,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"submodel","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.1,"co":0.5,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"submodel","id":"zai-org/GLM-4.5-FP8","n":"GLM 4.5 FP8","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.8,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"submodel","id":"zai-org/GLM-4.5-Air","n":"GLM 4.5 Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.1,"co":0.5,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"submodel","id":"deepseek-ai/DeepSeek-V3-0324","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2,"co":0.8,"ctx":75000,"out":163840,"mi":"t","mo":"t","tc":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"submodel","id":"deepseek-ai/DeepSeek-R1-0528","n":"DeepSeek R1 0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.15,"ctx":75000,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"submodel","id":"deepseek-ai/DeepSeek-V3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2,"co":0.8,"ctx":75000,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-23","lu":"2025-08-23"},{"p":"huggingface","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":0.59,"co":0.79,"ctx":131072,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"huggingface","id":"moonshotai/Kimi-K2-Thinking","n":"Kimi-K2-Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"huggingface","id":"moonshotai/Kimi-K2-Instruct-0905","n":"Kimi-K2-Instruct-0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":1,"co":3,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-04","lu":"2025-09-04"},{"p":"huggingface","id":"moonshotai/Kimi-K2-Instruct","n":"Kimi-K2-Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":1,"co":3,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-07-14","lu":"2025-07-14"},{"p":"huggingface","id":"moonshotai/Kimi-K2.6","n":"Kimi-K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-20","lu":"2026-04-20"},{"p":"huggingface","id":"moonshotai/Kimi-K2.5","n":"Kimi-K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-01","lu":"2026-01-01"},{"p":"huggingface","id":"moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"huggingface","id":"stepfun-ai/Step-3.5-Flash","n":"Step 3.5 Flash","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.1,"co":0.3,"ctx":262144,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-02-13"},{"p":"huggingface","id":"stepfun-ai/Step-3.7-Flash","n":"Step 3.7 Flash","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0.2,"co":1.15,"ctx":262144,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01-01","rd":"2026-05-29","lu":"2026-05-29"},{"p":"huggingface","id":"google/gemma-4-31B-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.14,"co":0.4,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"huggingface","id":"google/gemma-4-26B-A4B-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.13,"co":0.4,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"huggingface","id":"Qwen/Qwen3.6-35B-A3B","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.15,"co":0.95,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"huggingface","id":"Qwen/Qwen3-Coder-Next","n":"Qwen3-Coder-Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.2,"co":1.5,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-03","lu":"2026-02-03"},{"p":"huggingface","id":"Qwen/Qwen3-Embedding-8B","n":"Qwen 3 Embedding 8B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.01,"co":0,"ctx":32000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"huggingface","id":"Qwen/Qwen3.6-27B","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.47,"co":3.19,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"huggingface","id":"Qwen/Qwen3-Next-80B-A3B-Instruct","n":"Qwen3-Next-80B-A3B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.25,"co":1,"ctx":262144,"out":66536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-11","lu":"2025-09-11"},{"p":"huggingface","id":"Qwen/Qwen3.5-397B-A17B","n":"Qwen3.5-397B-A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-01","lu":"2026-02-01"},{"p":"huggingface","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen3-235B-A22B-Thinking-2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.3,"co":3,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-25","lu":"2025-07-25"},{"p":"huggingface","id":"Qwen/Qwen3-Embedding-4B","n":"Qwen 3 Embedding 4B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.01,"co":0,"ctx":32000,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"huggingface","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","n":"Qwen3-Coder-480B-A35B-Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":2,"co":2,"ctx":262144,"out":66536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"huggingface","id":"Qwen/Qwen3.5-122B-A10B","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":3.2,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"huggingface","id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","n":"Qwen3-Coder 30B-A3B Instruct","f":"qwen","d":"Smaller Qwen coder for efficient local agents and repo-level fixes","ci":0.07,"co":0.26,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"huggingface","id":"Qwen/Qwen3.5-27B","n":"Qwen3.5 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":2.4,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"huggingface","id":"Qwen/Qwen3-Next-80B-A3B-Thinking","n":"Qwen3-Next-80B-A3B-Thinking","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.3,"co":2,"ctx":262144,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-11","lu":"2025-09-11"},{"p":"huggingface","id":"Qwen/Qwen3.5-9B","n":"Qwen3.5 9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.17,"co":0.25,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"huggingface","id":"Qwen/Qwen3-32B","n":"Qwen3 32B","f":"qwen","d":"Dense open Qwen model for self-hosted chat, reasoning, and coding","ci":0.29,"co":0.59,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"huggingface","id":"Qwen/Qwen3-235B-A22B","n":"Qwen3 235B-A22B","f":"qwen","d":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","ci":0.2,"co":0.8,"ctx":40960,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"huggingface","id":"Qwen/Qwen3.5-35B-A3B","n":"Qwen3.5 35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.25,"co":2,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"huggingface","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.25,"co":0.69,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"huggingface","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.1,"co":0.5,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"huggingface","id":"XiaomiMiMo/MiMo-V2.5-Pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":1,"co":3,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"huggingface","id":"XiaomiMiMo/MiMo-V2-Flash","n":"MiMo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"ctx":262144,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-12-16","lu":"2025-12-16"},{"p":"huggingface","id":"zai-org/GLM-4.7-Flash","n":"GLM-4.7-Flash","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-08","lu":"2025-08-08"},{"p":"huggingface","id":"zai-org/GLM-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.55,"co":2.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"huggingface","id":"zai-org/GLM-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"huggingface","id":"zai-org/GLM-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.11,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"huggingface","id":"zai-org/GLM-4.5V","n":"GLM-4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"ctx":65536,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-11","lu":"2025-08-11"},{"p":"huggingface","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"huggingface","id":"zai-org/GLM-4.5-Air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.13,"co":0.85,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"huggingface","id":"zai-org/GLM-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-03","lu":"2026-04-03"},{"p":"huggingface","id":"zai-org/GLM-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"huggingface","id":"deepseek-ai/DeepSeek-R1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","ci":0.7,"co":2.5,"ctx":64000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"huggingface","id":"deepseek-ai/DeepSeek-R1-0528","n":"DeepSeek-R1-0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":3,"co":5,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2025-05-28","lu":"2025-05-28"},{"p":"huggingface","id":"deepseek-ai/DeepSeek-V4-Flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"huggingface","id":"deepseek-ai/DeepSeek-V4-Pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1048576,"out":393216,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"huggingface","id":"deepseek-ai/DeepSeek-V3.2","n":"DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":0.4,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"huggingface","id":"MiniMaxAI/MiniMax-M2.1","n":"MiniMax-M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-10","rd":"2025-12-23","lu":"2025-12-23"},{"p":"huggingface","id":"MiniMaxAI/MiniMax-M2","n":"MiniMax-M2","f":"minimax","d":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","ci":0.3,"co":1.2,"ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"huggingface","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"huggingface","id":"MiniMaxAI/MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.3,"co":1.2,"ctx":524288,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"huggingface","id":"MiniMaxAI/MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"minimax-coding-plan","id":"MiniMax-M2.1","n":"MiniMax-M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"minimax-coding-plan","id":"MiniMax-M2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"minimax-coding-plan","id":"MiniMax-M2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"minimax-coding-plan","id":"MiniMax-M2","n":"MiniMax-M2","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"ctx":196608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"minimax-coding-plan","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"minimax-coding-plan","id":"MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-25"},{"p":"minimax-coding-plan","id":"MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"novita-ai","id":"inclusionai/ling-2.6-1t","n":"Ling-2.6-1T","f":"ling","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.3,"co":2.5,"cr":0.06,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-04-23","lu":"2026-06-29"},{"p":"novita-ai","id":"inclusionai/ring-2.6-1t","n":"Ring-2.6-1T","f":"ring","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.3,"co":2.5,"cr":0.06,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-08","lu":"2026-05-27"},{"p":"novita-ai","id":"inclusionai/ling-2.6-flash","n":"Ling-2.6-flash","f":"ling","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.1,"co":0.3,"cr":0.02,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"novita-ai","id":"meta-llama/llama-3.1-8b-instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.02,"co":0.05,"ctx":16384,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2024-07-24","lu":"2024-07-24"},{"p":"novita-ai","id":"meta-llama/llama-3-70b-instruct","n":"Llama3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.51,"co":0.74,"ctx":8192,"out":8000,"mi":"t","mo":"t","ow":1,"rd":"2024-04-25","lu":"2024-04-25"},{"p":"novita-ai","id":"meta-llama/llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout Instruct","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.18,"co":0.59,"ctx":131072,"out":131072,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-06","lu":"2025-04-06"},{"p":"novita-ai","id":"meta-llama/llama-3.3-70b-instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.135,"co":0.4,"ctx":131072,"out":120000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-07","lu":"2024-12-07"},{"p":"novita-ai","id":"meta-llama/llama-3-8b-instruct","n":"Llama 3 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.04,"co":0.04,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-04-25","lu":"2024-04-25"},{"p":"novita-ai","id":"meta-llama/llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.03,"co":0.05,"ctx":32768,"out":32000,"mi":"t","mo":"t","ow":1,"rd":"2024-09-18","lu":"2024-09-18"},{"p":"novita-ai","id":"meta-llama/llama-4-maverick-17b-128e-instruct-fp8","n":"Llama 4 Maverick Instruct","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.27,"co":0.85,"ctx":1048576,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-06","lu":"2025-04-06"},{"p":"novita-ai","id":"moonshotai/kimi-k2-instruct","n":"Kimi K2 Instruct","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.57,"co":2.3,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-11","lu":"2025-07-11"},{"p":"novita-ai","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-07","lu":"2026-06-29"},{"p":"novita-ai","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"novita-ai","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.8,"co":3.4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"novita-ai","id":"moonshotai/kimi-k2-0905","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"novita-ai","id":"minimaxai/minimax-m1-80k","n":"MiniMax M1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.55,"co":2.2,"ctx":1000000,"out":40000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-17","lu":"2025-06-17"},{"p":"novita-ai","id":"baidu/ernie-4.5-vl-28b-a3b","n":"ERNIE 4.5 VL 28B A3B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.14,"co":0.56,"ctx":30000,"out":8000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-30","lu":"2026-06-14"},{"p":"novita-ai","id":"baidu/ernie-4.5-21B-a3b","n":"ERNIE 4.5 21B A3B","f":"ernie","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.07,"co":0.28,"ctx":120000,"out":8000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-06-30","lu":"2025-06-30"},{"p":"novita-ai","id":"baidu/ernie-4.5-vl-424b-a47b","n":"ERNIE 4.5 VL 424B A47B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.42,"co":1.25,"ctx":123000,"out":16000,"mi":"ti","mo":"t","rs":1,"ow":1,"rd":"2025-06-30","lu":"2025-06-30"},{"p":"novita-ai","id":"baidu/ernie-4.5-21B-a3b-thinking","n":"ERNIE-4.5-21B-A3B-Thinking","f":"ernie","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.07,"co":0.28,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2025-03","rd":"2025-09-19","lu":"2025-09-19"},{"p":"novita-ai","id":"baidu/ernie-4.5-300b-a47b-paddle","n":"ERNIE 4.5 300B A47B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.28,"co":1.1,"ctx":123000,"out":12000,"mi":"t","mo":"t","ow":1,"rd":"2025-06-30","lu":"2025-06-30"},{"p":"novita-ai","id":"baidu/ernie-4.5-vl-28b-a3b-thinking","n":"ERNIE-4.5-VL-28B-A3B-Thinking","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.39,"co":0.39,"ctx":131072,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-26","lu":"2025-11-26"},{"p":"novita-ai","id":"google/gemma-4-31b-it","n":"Gemma 4 31B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.14,"co":0.4,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"novita-ai","id":"google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.13,"co":0.4,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"novita-ai","id":"google/gemma-3-12b-it","n":"Gemma 3 12B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.05,"co":0.1,"ctx":131072,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-03-13","lu":"2025-03-13"},{"p":"novita-ai","id":"google/gemma-3-27b-it","n":"Gemma 3 27B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.119,"co":0.2,"ctx":98304,"out":16384,"mi":"ti","mo":"t","ow":1,"rd":"2025-03-25","lu":"2025-03-25"},{"p":"novita-ai","id":"microsoft/wizardlm-2-8x22b","n":"Wizardlm 2 8x22B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.62,"co":0.62,"ctx":65535,"out":8000,"mi":"t","mo":"t","ow":1,"rd":"2024-04-24","lu":"2024-04-24"},{"p":"novita-ai","id":"openai/gpt-oss-120b","n":"OpenAI GPT OSS 120B","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.25,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"novita-ai","id":"openai/gpt-oss-20b","n":"OpenAI: GPT OSS 20B","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.04,"co":0.15,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"ow":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"novita-ai","id":"sao10K/l31-70b-euryale-v2.2","n":"L31 70B Euryale V2.2","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":1.48,"co":1.48,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09-19","lu":"2024-09-19"},{"p":"novita-ai","id":"sao10K/L3-8B-stheno-v3.2","n":"L3 8B Stheno V3.2","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.05,"co":0.05,"ctx":8192,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-29","lu":"2024-11-29"},{"p":"novita-ai","id":"sao10K/l3-8b-lunaris","n":"Sao10k L3 8B Lunaris\t","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.05,"co":0.05,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-11-28","lu":"2024-11-28"},{"p":"novita-ai","id":"sao10K/l3-70b-euryale-v2.1","n":"L3 70B Euryale V2.1\t","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":1.48,"co":1.48,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-06-18","lu":"2024-06-18"},{"p":"novita-ai","id":"baichuan/baichuan-m2-32b","n":"baichuan-m2-32b","f":"baichuan","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.07,"co":0.07,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"kn":"2024-12","rd":"2025-08-13","lu":"2025-08-13"},{"p":"novita-ai","id":"mistralai/mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.04,"co":0.17,"ctx":60288,"out":16000,"mi":"t","mo":"t","ow":1,"rd":"2024-07-30","lu":"2024-07-30"},{"p":"novita-ai","id":"xiaomimimo/mimo-v2-flash","n":"XiaomiMiMo/MiMo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"cr":0.3,"ctx":262144,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-12-19","lu":"2025-12-19"},{"p":"novita-ai","id":"xiaomimimo/mimo-v2-pro","n":"MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":2,"co":6,"cr":0.4,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-05-27"},{"p":"novita-ai","id":"xiaomimimo/mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.522,"co":1.044,"cr":0.0043,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-27"},{"p":"novita-ai","id":"gryphe/mythomax-l2-13b","n":"Mythomax L2 13B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.09,"co":0.09,"ctx":4096,"out":3200,"mi":"t","mo":"t","ow":1,"rd":"2024-04-25","lu":"2024-04-25"},{"p":"novita-ai","id":"zai-org/glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.11,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-22","lu":"2025-12-22"},{"p":"novita-ai","id":"zai-org/glm-4.5v","n":"GLM 4.5V","f":"glmv","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"cr":0.11,"ctx":65536,"out":16384,"mi":"tvi","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-11","lu":"2025-08-11"},{"p":"novita-ai","id":"zai-org/glm-4.5","n":"GLM-4.5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.11,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"novita-ai","id":"zai-org/autoglm-phone-9b-multilingual","n":"AutoGLM-Phone-9B-Multilingual","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.035,"co":0.138,"ctx":65536,"out":65536,"mi":"ti","mo":"t","ow":1,"rd":"2025-12-10","lu":"2025-12-10"},{"p":"novita-ai","id":"zai-org/glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.38,"co":4.4,"cr":0.26,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"novita-ai","id":"zai-org/glm-4.6","n":"GLM 4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.55,"co":2.2,"cr":0.11,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-30","lu":"2025-09-30"},{"p":"novita-ai","id":"zai-org/glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"novita-ai","id":"zai-org/glm-4.6v","n":"GLM 4.6V","f":"glmv","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"cr":0.055,"ctx":131072,"out":32768,"mi":"tvi","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"novita-ai","id":"zai-org/glm-4.5-air","n":"GLM 4.5 Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.13,"co":0.85,"cr":0.025,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-10-13","lu":"2025-10-13"},{"p":"novita-ai","id":"zai-org/glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"cr":0.01,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"novita-ai","id":"zai-org/glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":202800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-12"},{"p":"novita-ai","id":"paddlepaddle/paddleocr-vl","n":"PaddleOCR-VL","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.02,"co":0.02,"ctx":16384,"out":16384,"mi":"ti","mo":"t","ow":1,"rd":"2025-10-22","lu":"2025-10-22"},{"p":"novita-ai","id":"qwen/qwen3-vl-235b-a22b-thinking","n":"Qwen3 VL 235B A22B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.98,"co":3.95,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","rs":1,"ow":1,"rd":"2025-09-24","lu":"2025-09-24"},{"p":"novita-ai","id":"qwen/qwen3-vl-30b-a3b-thinking","n":"qwen/qwen3-vl-30b-a3b-thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","tc":1,"ow":1,"rd":"2025-10-11","lu":"2025-10-11"},{"p":"novita-ai","id":"qwen/qwen3-235b-a22b-instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.58,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-22","lu":"2025-07-22"},{"p":"novita-ai","id":"qwen/qwen3-coder-30b-a3b-instruct","n":"Qwen3 Coder 30b A3B Instruct","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.07,"co":0.27,"ctx":160000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-10-09","lu":"2025-10-09"},{"p":"novita-ai","id":"qwen/qwen3-8b-fp8","n":"Qwen3 8B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.035,"co":0.138,"ctx":128000,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"novita-ai","id":"qwen/qwen3-next-80b-a3b-instruct","n":"Qwen3 Next 80B A3B Instruct","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":1.5,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-09-10","lu":"2025-09-10"},{"p":"novita-ai","id":"qwen/qwen3-vl-8b-instruct","n":"qwen/qwen3-vl-8b-instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.08,"co":0.5,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","tc":1,"ow":1,"rd":"2025-10-17","lu":"2025-10-17"},{"p":"novita-ai","id":"qwen/qwen3-omni-30b-a3b-thinking","n":"Qwen3 Omni 30B A3B Thinking","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.25,"co":0.97,"ctx":65536,"out":16384,"mi":"tavi","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-24","lu":"2025-09-24"},{"p":"novita-ai","id":"qwen/qwen3.7-max","n":"Qwen3.7-Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.25,"co":3.75,"cr":0.25,"cw":1.5625,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-27"},{"p":"novita-ai","id":"qwen/qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":2.11,"co":8.45,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-24","lu":"2025-09-24"},{"p":"novita-ai","id":"qwen/qwen2.5-7b-instruct","n":"Qwen2.5 7B Instruct","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.07,"co":0.07,"ctx":32000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"novita-ai","id":"qwen/qwen3-next-80b-a3b-thinking","n":"Qwen3 Next 80B A3B Thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.15,"co":1.5,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-10","lu":"2025-09-10"},{"p":"novita-ai","id":"qwen/qwen3-235b-a22b-thinking-2507","n":"Qwen3 235B A22b Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.3,"co":3,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-25","lu":"2025-07-25"},{"p":"novita-ai","id":"qwen/qwen-mt-plus","n":"Qwen MT Plus","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":0.25,"co":0.75,"ctx":16384,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2025-09-03","lu":"2025-09-03"},{"p":"novita-ai","id":"qwen/qwen3-32b-fp8","n":"Qwen3 32B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.45,"ctx":40960,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"novita-ai","id":"qwen/qwen3-4b-fp8","n":"Qwen3 4B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.03,"co":0.03,"ctx":128000,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"novita-ai","id":"qwen/qwen2.5-vl-72b-instruct","n":"Qwen2.5 VL 72B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.8,"co":0.8,"ctx":32768,"out":32768,"mi":"tiv","mo":"t","ow":1,"rd":"2025-03-25","lu":"2025-03-25"},{"p":"novita-ai","id":"qwen/qwen3-30b-a3b-fp8","n":"Qwen3 30B A3B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.45,"ctx":40960,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"novita-ai","id":"qwen/qwen3.5-27b","n":"Qwen3.5-27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":2.4,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-02-26"},{"p":"novita-ai","id":"qwen/qwen-2.5-72b-instruct","n":"Qwen 2.5 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.38,"co":0.4,"ctx":32000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-10-15","lu":"2024-10-15"},{"p":"novita-ai","id":"qwen/qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.2,"co":1.5,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-03","lu":"2026-02-03"},{"p":"novita-ai","id":"qwen/qwen3.5-35b-a3b","n":"Qwen3.5-35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.25,"co":2,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-02-26"},{"p":"novita-ai","id":"qwen/qwen3-coder-480b-a35b-instruct","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.38,"co":1.55,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"novita-ai","id":"qwen/qwen3.5-397b-a17b","n":"Qwen3.5-397B-A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":262144,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-17","lu":"2026-02-17"},{"p":"novita-ai","id":"qwen/qwen3-vl-30b-a3b-instruct","n":"qwen/qwen3-vl-30b-a3b-instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.7,"ctx":131072,"out":32768,"mi":"tvi","mo":"t","tc":1,"ow":1,"rd":"2025-10-11","lu":"2025-10-11"},{"p":"novita-ai","id":"qwen/qwen3-omni-30b-a3b-instruct","n":"Qwen3 Omni 30B A3B Instruct","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.25,"co":0.97,"ctx":65536,"out":16384,"mi":"tvai","mo":"ta","tc":1,"ow":1,"kn":"2024-04","rd":"2025-09-24","lu":"2025-09-24"},{"p":"novita-ai","id":"qwen/qwen3-vl-235b-a22b-instruct","n":"Qwen3 VL 235B A22B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.5,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","tc":1,"ow":1,"rd":"2025-09-24","lu":"2025-09-24"},{"p":"novita-ai","id":"qwen/qwen3.5-122b-a10b","n":"Qwen3.5-122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":3.2,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-02-26"},{"p":"novita-ai","id":"qwen/qwen3-235b-a22b-fp8","n":"Qwen3 235B A22B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.2,"co":0.8,"ctx":40960,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"novita-ai","id":"deepseek/deepseek-r1-0528","n":"DeepSeek R1 0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.7,"co":2.5,"cr":0.35,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-05-28","lu":"2025-05-28"},{"p":"novita-ai","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.028,"ctx":1048576,"out":393216,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"novita-ai","id":"deepseek/deepseek-v3.1-terminus","n":"Deepseek V3.1 Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"cr":0.135,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-22","lu":"2025-09-22"},{"p":"novita-ai","id":"deepseek/deepseek-v3-0324","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1.12,"cr":0.135,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2025-03-25","lu":"2025-03-25"},{"p":"novita-ai","id":"deepseek/deepseek-ocr","n":"DeepSeek-OCR","d":"OCR model for extracting structured text from documents and screenshots","ci":0.03,"co":0.03,"ctx":8192,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-10-24","lu":"2025-10-24"},{"p":"novita-ai","id":"deepseek/deepseek-r1-distill-llama-70b","n":"DeepSeek R1 Distill LLama 70B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.8,"co":0.8,"ctx":8192,"out":8192,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-01-27","lu":"2025-01-27"},{"p":"novita-ai","id":"deepseek/deepseek-r1-turbo","n":"DeepSeek R1 (Turbo)\t","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.7,"co":2.5,"ctx":64000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-03-05","lu":"2025-03-05"},{"p":"novita-ai","id":"deepseek/deepseek-prover-v2-671b","n":"Deepseek Prover V2 671B","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.7,"co":2.5,"ctx":160000,"out":160000,"mi":"t","mo":"t","ow":1,"rd":"2025-04-30","lu":"2025-04-30"},{"p":"novita-ai","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.6,"co":3.2,"cr":0.135,"ctx":1048576,"out":393216,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"novita-ai","id":"deepseek/deepseek-v3.2-exp","n":"Deepseek V3.2 Exp","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.41,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"novita-ai","id":"deepseek/deepseek-ocr-2","n":"deepseek/deepseek-ocr-2","d":"OCR model for extracting structured text from documents and screenshots","ci":0.03,"co":0.03,"ctx":8192,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"novita-ai","id":"deepseek/deepseek-r1-distill-qwen-14b","n":"DeepSeek R1 Distill Qwen 14B","f":"deepseek-thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.15,"co":0.15,"ctx":32768,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"novita-ai","id":"deepseek/deepseek-v3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"cr":0.135,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-21","lu":"2025-08-21"},{"p":"novita-ai","id":"deepseek/deepseek-r1-0528-qwen3-8b","n":"DeepSeek R1 0528 Qwen3 8B","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.06,"co":0.09,"ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-05-29","lu":"2025-05-29"},{"p":"novita-ai","id":"deepseek/deepseek-r1-distill-qwen-32b","n":"DeepSeek R1 Distill Qwen 32B","f":"deepseek-thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.3,"co":0.3,"ctx":64000,"out":32000,"mi":"t","mo":"t","ow":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"novita-ai","id":"deepseek/deepseek-v3-turbo","n":"DeepSeek V3 (Turbo)\t","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.4,"co":1.3,"ctx":64000,"out":16000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-03-05","lu":"2025-03-05"},{"p":"novita-ai","id":"deepseek/deepseek-v3.2","n":"Deepseek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.269,"co":0.4,"cr":0.1345,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"novita-ai","id":"minimax/minimax-m2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-05-27"},{"p":"novita-ai","id":"minimax/minimax-m2.5","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":204800,"out":131100,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"novita-ai","id":"minimax/minimax-m2.1","n":"Minimax M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":204800,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"novita-ai","id":"minimax/minimax-m2","n":"MiniMax-M2","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"novita-ai","id":"minimax/minimax-m2.7","n":"MiniMax M2.7","f":"minimax-m2.7","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"novita-ai","id":"minimax/minimax-m2.5-highspeed","n":"MiniMax M2.5 Highspeed","f":"minimax-m2.5","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.03,"ctx":204800,"out":131100,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"novita-ai","id":"kwaipilot/kat-coder-pro","n":"Kat Coder Pro","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":256000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01-05","lu":"2026-01-05"},{"p":"novita-ai","id":"nousresearch/hermes-2-pro-llama-3-8b","n":"Hermes 2 Pro Llama 3 8B","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.14,"co":0.14,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-06-27","lu":"2024-06-27"},{"p":"xai","id":"grok-4.20-multi-agent-0309","n":"Grok 4.20 Multi-Agent","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"xai","id":"grok-4.20-0309-non-reasoning","n":"Grok 4.20 (Non-Reasoning)","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"xai","id":"grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"xai","id":"grok-imagine-image-quality","n":"Grok Imagine Image Quality","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":8000,"out":0,"mi":"tip","mo":"ip","rd":"2026-04-03","lu":"2026-04-03"},{"p":"xai","id":"grok-imagine-video","n":"Grok Imagine Video","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":1024,"out":0,"mi":"tivp","mo":"v","rd":"2026-01-28","lu":"2026-01-28"},{"p":"xai","id":"grok-4.20-0309-reasoning","n":"Grok 4.20 (Reasoning)","f":"grok","d":"Reasoning Grok for document-heavy analysis and long-horizon tool use","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"xai","id":"grok-imagine-image","n":"Grok Imagine Image","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":8000,"out":0,"mi":"tip","mo":"ip","rd":"2026-01-28","lu":"2026-01-28"},{"p":"xai","id":"grok-build-0.1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"privatemode-ai","id":"qwen3-embedding-4b","n":"Qwen3-Embedding 4B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0,"co":0,"ctx":32000,"out":2560,"mi":"t","mo":"t","ow":1,"kn":"2025-06","rd":"2025-06-06","lu":"2025-06-06"},{"p":"privatemode-ai","id":"gemma-3-27b","n":"Gemma 3 27B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-03-12","lu":"2025-03-12"},{"p":"privatemode-ai","id":"gpt-oss-120b","n":"gpt-oss-120b","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0,"co":0,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-04","lu":"2025-08-14"},{"p":"privatemode-ai","id":"whisper-large-v3","n":"Whisper large-v3","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"a","mo":"t","ow":1,"kn":"2023-09","rd":"2023-09-01","lu":"2023-09-01"},{"p":"privatemode-ai","id":"qwen3-coder-30b-a3b","n":"Qwen3-Coder 30B-A3B","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"drun","id":"public/deepseek-v3","n":"DeepSeek V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":1.1,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-12-26","lu":"2024-12-26"},{"p":"drun","id":"public/deepseek-r1","n":"DeepSeek R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.55,"co":2.2,"ctx":131072,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-20","lu":"2025-01-20"},{"p":"drun","id":"public/minimax-m25","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.29,"co":1.16,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-03-01","lu":"2025-03-01"},{"p":"alibaba-token-plan-cn","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0,"co":0,"cr":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"alibaba-token-plan-cn","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"alibaba-token-plan-cn","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"alibaba-token-plan-cn","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"alibaba-token-plan-cn","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"alibaba-token-plan-cn","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0,"co":0,"cr":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"alibaba-token-plan-cn","id":"wan2.7-image-pro","n":"Wan2.7 Image Pro","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-05-29","lu":"2026-05-29"},{"p":"alibaba-token-plan-cn","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"alibaba-token-plan-cn","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"alibaba-token-plan-cn","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"alibaba-token-plan-cn","id":"qwen-image-2.0","n":"Qwen Image 2.0","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-03-03","lu":"2026-03-03"},{"p":"alibaba-token-plan-cn","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0,"co":0,"cr":0,"cw":0,"ctx":196608,"out":24576,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-token-plan-cn","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"alibaba-token-plan-cn","id":"qwen-image-2.0-pro","n":"Qwen Image 2.0 Pro","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-03-03","lu":"2026-03-03"},{"p":"alibaba-token-plan-cn","id":"wan2.7-image","n":"Wan2.7 Image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-05-29","lu":"2026-05-29"},{"p":"alibaba-token-plan-cn","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-token-plan-cn","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0,"co":0,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-12-03","lu":"2025-12-05"},{"p":"alibaba-token-plan-cn","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"moonshotai","id":"kimi-k2-0905-preview","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"moonshotai","id":"kimi-k2-thinking-turbo","n":"Kimi K2 Thinking Turbo","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":1.15,"co":8,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"moonshotai","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"moonshotai","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Thinking Kimi model for slower research passes, planning, and hard technical questions","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"moonshotai","id":"kimi-k2-0711-preview","n":"Kimi K2 0711","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"cr":0.15,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-07-14","lu":"2025-07-14"},{"p":"moonshotai","id":"kimi-k2-turbo-preview","n":"Kimi K2 Turbo","f":"kimi-k2","d":"Fast Kimi model for responsive chat, coding help, and agent loops","ci":2.4,"co":10,"cr":0.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"moonshotai","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"moonshotai","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"moonshotai","id":"kimi-k2.7-code-highspeed","n":"Kimi K2.7 Code HighSpeed","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":1.9,"co":8,"cr":0.38,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"fireworks-ai","id":"accounts/fireworks/routers/kimi-k2p6-turbo","n":"Kimi K2.6 Turbo","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":2,"co":8,"cr":0.3,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"fireworks-ai","id":"accounts/fireworks/routers/glm-5p2-fast","n":"GLM 5.2 Fast","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":2.1,"co":6.6,"cr":0.21,"ctx":1048575,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-26","lu":"2026-06-26"},{"p":"fireworks-ai","id":"accounts/fireworks/routers/kimi-k2p7-code-fast","n":"Kimi K2.7 Code Fast","f":"kimi-k2","d":"Kimi coding model for software agents, refactors, and repository reasoning","ci":1.9,"co":8,"cr":0.38,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-12","lu":"2026-06-16"},{"p":"fireworks-ai","id":"accounts/fireworks/routers/glm-5p1-fast","n":"GLM 5.1 Fast","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":2.8,"co":8.8,"cr":0.52,"ctx":202800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"fireworks-ai","id":"accounts/fireworks/routers/kimi-k2p6-fast","n":"Kimi K2.6 Fast","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":2,"co":8,"cr":0.3,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-06-05"},{"p":"fireworks-ai","id":"accounts/fireworks/models/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-06-16"},{"p":"fireworks-ai","id":"accounts/fireworks/models/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.74,"co":3.48,"cr":0.145,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"fireworks-ai","id":"accounts/fireworks/models/minimax-m2p7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-12","lu":"2026-04-12"},{"p":"fireworks-ai","id":"accounts/fireworks/models/minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":512000,"out":512000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-12","lu":"2026-06-12"},{"p":"fireworks-ai","id":"accounts/fireworks/models/kimi-k2p6","n":"Kimi K2.6","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.95,"co":4,"cr":0.16,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"fireworks-ai","id":"accounts/fireworks/models/qwen3p7-plus","n":"Qwen 3.7 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":1.6,"cr":0.08,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-12","lu":"2026-06-12"},{"p":"fireworks-ai","id":"accounts/fireworks/models/glm-5p1","n":"GLM 5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cr":0.26,"ctx":202800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"fireworks-ai","id":"accounts/fireworks/models/glm-5p2","n":"GLM 5.2","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cr":0.26,"ctx":1048575,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-16","lu":"2026-06-16"},{"p":"fireworks-ai","id":"accounts/fireworks/models/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"cr":0.015,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2026-06-16"},{"p":"fireworks-ai","id":"accounts/fireworks/models/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.07,"co":0.3,"cr":0.035,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"fireworks-ai","id":"accounts/fireworks/models/kimi-k2p7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Kimi coding model for software agents, refactors, and repository reasoning","ci":0.95,"co":4,"cr":0.19,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-12","lu":"2026-06-16"},{"p":"vultr","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.15,"co":0.6,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"vultr","id":"nvidia/Llama-3.1-Nemotron-Safety-Guard-8B-v3","n":"Llama 3.1 Nemotron Safety Guard","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.01,"co":0.01,"ctx":8192,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-12","rd":"2025-10-28","lu":"2025-10-28"},{"p":"vultr","id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16","n":"NVIDIA Nemotron 3 Nano Omni","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0.13,"co":0.38,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-28","lu":"2026-04-28"},{"p":"vultr","id":"nvidia/DeepSeek-V3.2-NVFP4","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.55,"co":1.65,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"vultr","id":"nvidia/Nemotron-Cascade-2-30B-A3B","n":"NVIDIA Nemotron Cascade 2","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.15,"co":0.6,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"vultr","id":"zai-org/GLM-5.1-FP8","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.85,"co":3.1,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"vultr","id":"MiniMaxAI/MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"302ai","id":"gpt-5.4-mini-2026-03-17","n":"gpt-5.4-mini-2026-03-17","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-19","lu":"2026-03-19"},{"p":"302ai","id":"chatgpt-4o-latest","n":"chatgpt-4o-latest","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":15,"ctx":128000,"out":16384,"mi":"ti","mo":"t","kn":"2023-09","rd":"2024-08-08","lu":"2024-08-08"},{"p":"302ai","id":"gpt-5.4-nano-2026-03-17","n":"gpt-5.4-nano-2026-03-17","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-19","lu":"2026-03-19"},{"p":"302ai","id":"kimi-k2-0905-preview","n":"kimi-k2-0905-preview","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.632,"co":2.53,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"kn":"2025-06","rd":"2025-09-05","lu":"2025-09-05"},{"p":"302ai","id":"grok-4.20-beta-0309-non-reasoning","n":"grok-4.20-beta-0309-non-reasoning","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"302ai","id":"gemini-2.5-flash-nothink","n":"gemini-2.5-flash-nothink","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-06-24","lu":"2025-06-24"},{"p":"302ai","id":"qwen-plus","n":"Qwen-Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.12,"co":1.2,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2024-07-23","lu":"2024-07-23"},{"p":"302ai","id":"glm-4.7","n":"glm-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.286,"co":1.142,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"302ai","id":"qwen3-235b-a22b-instruct-2507","n":"qwen3-235b-a22b-instruct-2507","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.29,"co":1.143,"ctx":128000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-30","lu":"2025-07-30"},{"p":"302ai","id":"glm-4.5v","n":"GLM-4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.29,"co":0.86,"ctx":64000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-12","lu":"2025-08-12"},{"p":"302ai","id":"claude-opus-4-5","n":"claude-opus-4-5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-25","lu":"2025-11-25"},{"p":"302ai","id":"gemini-2.5-pro","n":"gemini-2.5-pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"302ai","id":"gpt-5","n":"gpt-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-08","lu":"2025-08-08"},{"p":"302ai","id":"claude-haiku-4-5-20251001","n":"claude-haiku-4-5-20251001","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-16","lu":"2025-10-16"},{"p":"302ai","id":"kimi-k2-thinking-turbo","n":"kimi-k2-thinking-turbo","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":1.265,"co":9.119,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-09-05","lu":"2025-09-05"},{"p":"302ai","id":"claude-3-5-haiku-20241022","n":"claude-3-5-haiku-20241022","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"302ai","id":"glm-4.5","n":"GLM-4.5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.286,"co":1.142,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-29","lu":"2025-07-29"},{"p":"302ai","id":"gpt-5-pro","n":"gpt-5-pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-08","lu":"2025-10-08"},{"p":"302ai","id":"grok-4.20-beta-0309-reasoning","n":"grok-4.20-beta-0309-reasoning","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"302ai","id":"gemini-2.5-flash","n":"gemini-2.5-flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"302ai","id":"gpt-4o","n":"gpt-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-05-13"},{"p":"302ai","id":"MiniMax-M2.1","n":"MiniMax-M2.1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":1000000,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2025-12-19","lu":"2025-12-19"},{"p":"302ai","id":"gemini-2.5-flash-lite-preview-09-2025","n":"gemini-2.5-flash-lite-preview-09-2025","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-09-26","lu":"2025-09-26"},{"p":"302ai","id":"doubao-seed-1-6-vision-250815","n":"doubao-seed-1-6-vision-250815","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.114,"co":1.143,"ctx":256000,"out":32000,"mi":"ti","mo":"t","tc":1,"rd":"2025-09-30","lu":"2025-09-30"},{"p":"302ai","id":"claude-opus-4-1-20250805","n":"claude-opus-4-1-20250805","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"302ai","id":"qwen3-max-2025-09-23","n":"qwen3-max-2025-09-23","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.86,"co":3.43,"ctx":258048,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-24","lu":"2025-09-24"},{"p":"302ai","id":"glm-4.7-flashx","n":"glm-4.7-flashx","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.0715,"co":0.429,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-20","lu":"2026-01-20"},{"p":"302ai","id":"glm-5.1","n":"glm-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.86,"co":3.5,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-10","lu":"2026-04-10"},{"p":"302ai","id":"glm-4.6","n":"glm-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.286,"co":1.142,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"302ai","id":"kimi-k2-thinking","n":"kimi-k2-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.575,"co":2.3,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-09-05","lu":"2025-09-05"},{"p":"302ai","id":"claude-sonnet-4-5","n":"claude-sonnet-4-5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-30","lu":"2025-09-30"},{"p":"302ai","id":"glm-4.5-x","n":"glm-4.5-x","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.143,"co":2.29,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-29","lu":"2025-07-29"},{"p":"302ai","id":"deepseek-v3.2-thinking","n":"DeepSeek-V3.2-Thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.29,"co":0.43,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2025-12-01","lu":"2025-12-01"},{"p":"302ai","id":"claude-sonnet-4-6-thinking","n":"claude-sonnet-4-6-thinking","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08","rd":"2026-02-18","lu":"2026-03-13"},{"p":"302ai","id":"claude-opus-4-7","n":"claude-opus-4-7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-17","lu":"2026-04-17"},{"p":"302ai","id":"grok-4-1-fast-non-reasoning","n":"grok-4-1-fast-non-reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","tc":1,"kn":"2025-06","rd":"2025-11-20","lu":"2025-11-20"},{"p":"302ai","id":"gpt-5.4-nano","n":"gpt-5.4-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-19","lu":"2026-03-19"},{"p":"302ai","id":"claude-opus-4-1-20250805-thinking","n":"claude-opus-4-1-20250805-thinking","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2025-05-27","lu":"2025-05-27"},{"p":"302ai","id":"glm-4.5-airx","n":"glm-4.5-airx","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.572,"co":1.714,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-29","lu":"2025-07-29"},{"p":"302ai","id":"grok-4.1","n":"grok-4.1","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":10,"ctx":200000,"out":64000,"mi":"ti","mo":"t","tc":1,"kn":"2025-06","rd":"2025-11-18","lu":"2025-11-18"},{"p":"302ai","id":"gemini-2.5-flash-preview-09-2025","n":"gemini-2.5-flash-preview-09-2025","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-09-26","lu":"2025-09-26"},{"p":"302ai","id":"claude-opus-4-5-20251101","n":"claude-opus-4-5-20251101","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-25","lu":"2025-11-25"},{"p":"302ai","id":"claude-opus-4-20250514","n":"claude-opus-4-20250514","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"302ai","id":"gemini-3-pro-image-preview","n":"gemini-3-pro-image-preview","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":120,"ctx":32768,"out":64000,"mi":"ti","mo":"t","kn":"2025-06","rd":"2025-11-20","lu":"2025-11-20"},{"p":"302ai","id":"gemini-2.5-flash-image","n":"gemini-2.5-flash-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.3,"co":30,"ctx":32768,"out":32768,"mi":"ti","mo":"t","kn":"2025-01","rd":"2025-10-08","lu":"2025-10-08"},{"p":"302ai","id":"glm-for-coding","n":"glm-for-coding","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.086,"co":0.343,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-30","lu":"2025-09-30"},{"p":"302ai","id":"gpt-5.2","n":"gpt-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-12","lu":"2025-12-12"},{"p":"302ai","id":"qwen-flash","n":"Qwen-Flash","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.022,"co":0.22,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"302ai","id":"claude-opus-4-6-thinking","n":"claude-opus-4-6-thinking","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05","rd":"2026-02-06","lu":"2026-03-13"},{"p":"302ai","id":"claude-sonnet-4-20250514","n":"claude-sonnet-4-20250514","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"302ai","id":"gpt-5.1-chat-latest","n":"gpt-5.1-chat-latest","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"302ai","id":"gpt-5.2-chat-latest","n":"gpt-5.2-chat-latest","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-12","lu":"2025-12-12"},{"p":"302ai","id":"grok-4-fast-reasoning","n":"grok-4-fast-reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-09-23","lu":"2025-09-23"},{"p":"302ai","id":"gpt-4.1-nano","n":"gpt-4.1-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"302ai","id":"claude-sonnet-4-5-20250929-thinking","n":"claude-sonnet-4-5-20250929-thinking","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2025-09-30","lu":"2025-09-30"},{"p":"302ai","id":"MiniMax-M2.7-highspeed","n":"MiniMax-M2.7-highspeed","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":4.8,"ctx":204800,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2026-03-19","lu":"2026-03-19"},{"p":"302ai","id":"MiniMax-M2","n":"MiniMax-M2","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.33,"co":1.32,"ctx":1000000,"out":128000,"mi":"t","mo":"t","tc":1,"rd":"2025-10-26","lu":"2025-10-26"},{"p":"302ai","id":"gemini-3.1-flash-image-preview","n":"gemini-3.1-flash-image-preview","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":60,"ctx":131072,"out":32768,"mi":"tip","mo":"ti","kn":"2025-01","rd":"2026-02-27","lu":"2026-02-27"},{"p":"302ai","id":"qwen3-235b-a22b","n":"Qwen3-235B-A22B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.29,"co":2.86,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-04-29","lu":"2025-04-29"},{"p":"302ai","id":"ministral-14b-2512","n":"ministral-14b-2512","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.33,"co":0.33,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2024-12","rd":"2025-12-16","lu":"2025-12-16"},{"p":"302ai","id":"glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.145,"co":0.43,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"302ai","id":"claude-haiku-4-5","n":"claude-haiku-4-5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-16","lu":"2025-10-16"},{"p":"302ai","id":"gpt-5.4","n":"gpt-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"cw":0,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"302ai","id":"doubao-seed-1-6-thinking-250715","n":"doubao-seed-1-6-thinking-250715","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.121,"co":1.21,"ctx":256000,"out":16000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-07-15","lu":"2025-07-15"},{"p":"302ai","id":"gpt-5.4-mini","n":"gpt-5.4-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-19","lu":"2026-03-19"},{"p":"302ai","id":"gpt-4.1","n":"gpt-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"302ai","id":"doubao-seed-code-preview-251028","n":"doubao-seed-code-preview-251028","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.17,"co":1.14,"ctx":256000,"out":32000,"mi":"ti","mo":"t","tc":1,"rd":"2025-11-11","lu":"2025-11-11"},{"p":"302ai","id":"claude-opus-4-6","n":"claude-opus-4-6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-06","lu":"2026-03-13"},{"p":"302ai","id":"qwen3-coder-480b-a35b-instruct","n":"qwen3-coder-480b-a35b-instruct","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.86,"co":3.43,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"302ai","id":"claude-sonnet-4-5-20250929","n":"claude-sonnet-4-5-20250929","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-30","lu":"2025-09-30"},{"p":"302ai","id":"deepseek-reasoner","n":"Deepseek-Reasoner","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.29,"co":0.43,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-01-20"},{"p":"302ai","id":"grok-4-1-fast-reasoning","n":"grok-4-1-fast-reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-11-20","lu":"2025-11-20"},{"p":"302ai","id":"gemini-3-pro-preview","n":"gemini-3-pro-preview","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","tc":1,"kn":"2025-06","rd":"2025-11-19","lu":"2025-11-19"},{"p":"302ai","id":"gpt-5-thinking","n":"gpt-5-thinking","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2025-08-08","lu":"2025-08-08"},{"p":"302ai","id":"gpt-5-mini","n":"gpt-5-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-08","lu":"2025-08-08"},{"p":"302ai","id":"gpt-4.1-mini","n":"gpt-4.1-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"302ai","id":"glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.72,"co":3.2,"ctx":200000,"out":131072,"mi":"tivap","mo":"t","rs":1,"tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"302ai","id":"gpt-5.4-pro","n":"gpt-5.4-pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"cr":0,"cw":0,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"302ai","id":"glm-4.5-air","n":"glm-4.5-air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.1143,"co":0.286,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-29","lu":"2025-07-29"},{"p":"302ai","id":"claude-sonnet-4-6","n":"claude-sonnet-4-6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-18","lu":"2026-03-13"},{"p":"302ai","id":"grok-4-fast-non-reasoning","n":"grok-4-fast-non-reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","tc":1,"kn":"2025-06","rd":"2025-09-23","lu":"2025-09-23"},{"p":"302ai","id":"gemini-3-flash-preview","n":"gemini-3-flash-preview","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","tc":1,"kn":"2025-06","rd":"2025-12-18","lu":"2025-12-18"},{"p":"302ai","id":"deepseek-chat","n":"Deepseek-Chat","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.29,"co":0.43,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-11-29","lu":"2024-11-29"},{"p":"302ai","id":"MiniMax-M1","n":"MiniMax-M1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.132,"co":1.254,"ctx":1000000,"out":128000,"mi":"t","mo":"t","tc":1,"rd":"2025-06-16","lu":"2025-06-16"},{"p":"302ai","id":"grok-4.20-multi-agent-beta-0309","n":"grok-4.20-multi-agent-beta-0309","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"302ai","id":"glm-5","n":"glm-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.6,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"302ai","id":"qwen-max-latest","n":"Qwen-Max-Latest","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0.343,"co":1.372,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-11","rd":"2024-04-03","lu":"2025-01-25"},{"p":"302ai","id":"mistral-large-2512","n":"mistral-large-2512","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":1.1,"co":3.3,"ctx":128000,"out":262144,"mi":"ti","mo":"t","tc":1,"kn":"2024-12","rd":"2025-12-16","lu":"2025-12-16"},{"p":"302ai","id":"MiniMax-M2.7","n":"MiniMax-M2.7","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2026-03-19","lu":"2026-03-19"},{"p":"302ai","id":"claude-3-5-haiku-latest","n":"claude-3-5-haiku-latest","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"302ai","id":"qwen3-30b-a3b","n":"Qwen3-30B-A3B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.11,"co":1.08,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-04-29","lu":"2025-04-29"},{"p":"302ai","id":"deepseek-v3.2","n":"deepseek-v3.2","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.29,"co":0.43,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2025-12-01","lu":"2025-12-01"},{"p":"302ai","id":"doubao-seed-1-8-251215","n":"doubao-seed-1-8-251215","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.114,"co":0.286,"ctx":224000,"out":64000,"mi":"ti","mo":"t","tc":1,"rd":"2025-12-18","lu":"2025-12-18"},{"p":"302ai","id":"claude-opus-4-5-20251101-thinking","n":"claude-opus-4-5-20251101-thinking","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2025-11-25","lu":"2025-11-25"},{"p":"302ai","id":"gemini-2.0-flash-lite","n":"gemini-2.0-flash-lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.075,"co":0.3,"ctx":2000000,"out":8192,"mi":"ti","mo":"t","kn":"2024-11","rd":"2025-06-16","lu":"2025-06-16"},{"p":"302ai","id":"glm-5-turbo","n":"glm-5-turbo","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.72,"co":3.2,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"302ai","id":"gpt-5.1","n":"gpt-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"trustedrouter","id":"zdr","n":"Zero Data Retention","d":"TrustedRouter privacy routing alias that prefers zero data retention model endpoints.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-01","lu":"2026-06-27"},{"p":"trustedrouter","id":"e2e","n":"End-to-End Encrypted","d":"TrustedRouter privacy routing alias for end-to-end encrypted provider routes where available.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-01","lu":"2026-06-27"},{"p":"trustedrouter","id":"synth-code","n":"Synth Code","d":"TrustedRouter code synthesis orchestration alias that combines multiple model responses into one answer.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-20","lu":"2026-06-27"},{"p":"trustedrouter","id":"fast","n":"Fast","d":"TrustedRouter speed routing alias that prefers low-latency healthy model endpoints.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-01","lu":"2026-06-27"},{"p":"trustedrouter","id":"synth","n":"Synth","d":"TrustedRouter synthesis orchestration alias that combines multiple model responses into one answer.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-20","lu":"2026-06-27"},{"p":"trustedrouter","id":"auto","n":"Auto","d":"TrustedRouter automatic routing alias that chooses a healthy supported model endpoint for the request.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-01","lu":"2026-06-27"},{"p":"trustedrouter","id":"cheap","n":"Cheap","d":"TrustedRouter low-cost routing alias that prefers inexpensive healthy model endpoints.","ctx":1000000,"out":131072,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-01","lu":"2026-06-27"},{"p":"zhipuai","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"zhipuai","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"zhipuai","id":"glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":5,"co":22,"cr":1.2,"cw":0,"ctx":200000,"out":131072,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"zhipuai","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":1,"co":3.2,"cr":0.2,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"zhipuai","id":"glm-4.5-flash","n":"GLM-4.5-Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zhipuai","id":"glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"zhipuai","id":"glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.2,"co":1.1,"cr":0.03,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zhipuai","id":"glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"zhipuai","id":"glm-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"zhipuai","id":"glm-4.7-flashx","n":"GLM-4.7-FlashX","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"cr":0.01,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"zhipuai","id":"glm-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zhipuai","id":"glm-4.5v","n":"GLM-4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"ctx":64000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-11","lu":"2025-08-11"},{"p":"zhipuai","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"cortecs","id":"deepseek-r1-0528","n":"DeepSeek R1 0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.585,"co":2.307,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-05-28","lu":"2025-05-28"},{"p":"cortecs","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.133,"co":0.266,"cr":0.0028,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"cortecs","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.32,"co":1.18,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"cortecs","id":"deepseek-v3-0324","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.551,"co":1.654,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2025-03-24","lu":"2025-03-24"},{"p":"cortecs","id":"claude-opus4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5.6,"co":27.99,"cr":0.56,"cw":6.99,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"cortecs","id":"glm-4.7","n":"GLM 4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.45,"co":2.23,"ctx":198000,"out":198000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"cortecs","id":"qwen3-235b-a22b-instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.062,"co":0.408,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"cortecs","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3 Coder 30B A3B Instruct","f":"qwen","d":"Smaller Qwen coder for efficient local agents and repo-level fixes","ci":0.053,"co":0.222,"ctx":262000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-31","lu":"2025-07-31"},{"p":"cortecs","id":"minimax-m2.1","n":"MiniMax-M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.34,"co":1.34,"ctx":196000,"out":196000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"cortecs","id":"qwen3-32b","n":"Qwen3 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.099,"co":0.33,"ctx":16384,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-04-29","lu":"2025-04-29"},{"p":"cortecs","id":"claude-4-6-sonnet","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3.59,"co":17.92,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"cortecs","id":"claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3.307,"co":16.536,"ctx":200000,"out":64000,"mi":"tip","mo":"t","tc":1,"kn":"2025-03","rd":"2025-05-22","lu":"2025-05-22"},{"p":"cortecs","id":"llama-4-maverick","n":"Llama 4 Maverick 17B Instruct","f":"llama","d":"Open multimodal Llama for strong reasoning with efficient everyday serving","ci":0.124,"co":0.603,"cr":0.03,"cw":0.151,"ctx":1000000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"cortecs","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.654,"co":11.024,"ctx":1048576,"out":65535,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-17"},{"p":"cortecs","id":"claude-4-5-sonnet","n":"Claude 4.5 Sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3.259,"co":16.296,"ctx":200000,"out":200000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"cortecs","id":"mixtral-8x7B-instruct-v0.1","n":"Mixtral 8x7B Instruct v0.1","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.438,"co":0.68,"ctx":32000,"out":32000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2023-09","rd":"2023-12-11","lu":"2023-12-11"},{"p":"cortecs","id":"glm-4.5","n":"GLM 4.5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.67,"co":2.46,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-29","lu":"2025-07-29"},{"p":"cortecs","id":"kimi-k2-instruct","n":"Kimi K2 Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.551,"co":2.646,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2025-07-11","lu":"2025-09-05"},{"p":"cortecs","id":"llama-3.3-70b-instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":0.089,"co":0.275,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"cortecs","id":"claude-opus4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5.64,"co":28.198,"cr":0.563,"cw":7.049,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"cortecs","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":1.28,"co":4.63,"cr":0.32,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"cortecs","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.31,"co":4.1,"cr":0.24,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-14","lu":"2026-04-14"},{"p":"cortecs","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.553,"co":3.106,"cr":0.003625,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"cortecs","id":"qwen3-next-80b-a3b-thinking","n":"Qwen3 Next 80B A3B Thinking","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.164,"co":1.311,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-11","lu":"2025-09-11"},{"p":"cortecs","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.656,"co":2.731,"ctx":262000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-08","lu":"2025-12-08"},{"p":"cortecs","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.44,"co":4.53,"cr":0.39,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"cortecs","id":"minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.355,"co":1.775,"cr":0.089,"ctx":512000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"cortecs","id":"minimax-m2","n":"MiniMax-M2","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.39,"co":1.57,"ctx":400000,"out":400000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-11","rd":"2025-10-27","lu":"2025-10-27"},{"p":"cortecs","id":"minimax-m2.7","n":"MiniMax-m2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.47,"co":1.4,"ctx":202752,"out":196072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"cortecs","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.55,"co":2.76,"ctx":256000,"out":256000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"cortecs","id":"gpt-oss-120b","n":"GPT Oss 120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0,"co":0,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-01","rd":"2025-08-05","lu":"2025-08-05"},{"p":"cortecs","id":"qwen-2.5-72b-instruct","n":"Qwen2.5 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.062,"co":0.231,"ctx":33000,"out":33000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06","rd":"2024-09-19","lu":"2024-09-19"},{"p":"cortecs","id":"hermes-4-70b","n":"Hermes 4 70B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.116,"co":0.358,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2025-08-26","lu":"2025-08-26"},{"p":"cortecs","id":"claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1.09,"co":5.43,"ctx":200000,"out":200000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"cortecs","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":3,"co":16.13,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"cortecs","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.81,"co":3.54,"cr":0.2,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"cortecs","id":"gpt-4.1","n":"GPT 4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.354,"co":9.417,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-06","rd":"2025-04-14","lu":"2025-04-14"},{"p":"cortecs","id":"nova-pro-v1","n":"Nova Pro 1.0","f":"nova-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":1.016,"co":4.061,"ctx":300000,"out":5000,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-12-03","lu":"2024-12-03"},{"p":"cortecs","id":"qwen3-coder-next","n":"Qwen3 Coder Next 80B","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.158,"co":0.84,"ctx":256000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-04","lu":"2026-02-04"},{"p":"cortecs","id":"claude-opus4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5.98,"co":29.89,"ctx":200000,"out":200000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"cortecs","id":"qwen3-coder-480b-a35b-instruct","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.441,"co":1.984,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-01","rd":"2025-07-25","lu":"2025-07-25"},{"p":"cortecs","id":"devstral-2512","n":"Devstral 2 2512","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0,"co":0,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"cortecs","id":"qwen3.5-397b-a17b","n":"Qwen3.5 397B A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.6,"co":3.6,"ctx":250000,"out":250000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-02-16","lu":"2026-02-16"},{"p":"cortecs","id":"glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":1.235,"co":4.118,"cr":0.308,"cw":1.544,"ctx":200000,"out":131072,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"cortecs","id":"glm-4.5-air","n":"GLM 4.5 Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.22,"co":1.34,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-01","lu":"2025-08-01"},{"p":"cortecs","id":"glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.09,"co":0.53,"ctx":203000,"out":203000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-08","lu":"2025-08-08"},{"p":"cortecs","id":"intellect-3","n":"INTELLECT 3","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.219,"co":1.202,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-11","rd":"2025-11-26","lu":"2025-11-26"},{"p":"cortecs","id":"llama-3.1-405b-instruct","n":"Llama 3.1 405B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"cortecs","id":"glm-5","n":"GLM 5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.08,"co":3.44,"ctx":202752,"out":202752,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"cortecs","id":"mistral-large-2512","n":"Mistral Large 3 2512","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ci":0.5,"co":1.5,"cr":0.05,"ctx":256000,"out":256000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-01","lu":"2025-12-01"},{"p":"cortecs","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.266,"co":0.444,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"cortecs","id":"codestral-2508","n":"Codestral 2508","f":"mistral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.9,"cr":0.03,"ctx":256000,"out":256000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-07-30","lu":"2025-07-30"},{"p":"cortecs","id":"qwen3.5-122b-a10b","n":"Qwen3.5 122B A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.444,"co":3.106,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-02-24","lu":"2026-02-24"},{"p":"cortecs","id":"nemotron-3-super-120b-a12b","n":"Nemotron 3 Super 120B A12B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.266,"co":0.799,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-03-11","lu":"2026-03-11"},{"p":"cortecs","id":"glm-5-turbo","n":"GLM-5-Turbo","f":"glm","d":"Faster GLM-5 lane for coding agents that need lower latency","ci":1.235,"co":4.118,"cr":0.308,"cw":1.544,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"cortecs","id":"claude-opus4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5.98,"co":29.89,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"nebius","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama-3.3-70B-Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.13,"co":0.4,"cr":0.013,"cw":0.16,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-08","rd":"2025-12-05","lu":"2026-02-04"},{"p":"nebius","id":"moonshotai/Kimi-K2.5-fast","n":"Kimi-K2.5-fast","f":"kimi-k2","d":"Legacy model retained for compatibility with older integrations","ci":0.5,"co":2.5,"cr":0.05,"cw":0.625,"ctx":256000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-12-15","lu":"2026-02-04"},{"p":"nebius","id":"moonshotai/Kimi-K2.5","n":"Kimi-K2.5","f":"kimi-k2","d":"Legacy model retained for compatibility with older integrations","ci":0.5,"co":2.5,"cr":0.05,"cw":0.625,"ctx":256000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-12-15","lu":"2026-02-04"},{"p":"nebius","id":"google/gemma-3-27b-it","n":"Gemma-3-27b-it","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.1,"co":0.3,"cr":0.01,"cw":0.125,"ctx":110000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-10","rd":"2026-01-20","lu":"2026-02-04"},{"p":"nebius","id":"Qwen/Qwen3-Next-80B-A3B-Thinking-fast","n":"Qwen3-Next-80B-A3B-Thinking-fast","d":"Legacy model retained for compatibility with older integrations","ci":0.15,"co":1.2,"cr":0.015,"cw":0.1875,"ctx":8000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-25","lu":"2026-05-07"},{"p":"nebius","id":"Qwen/Qwen2.5-VL-72B-Instruct","n":"Qwen2.5-VL-72B-Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.25,"co":0.75,"cr":0.025,"cw":0.31,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-20","lu":"2026-02-04"},{"p":"nebius","id":"Qwen/Qwen3-Embedding-8B","n":"Qwen3-Embedding-8B","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.01,"co":0,"ctx":32768,"out":0,"mi":"t","mo":"t","ow":1,"kn":"2025-10","rd":"2026-01-10","lu":"2026-02-04"},{"p":"nebius","id":"Qwen/Qwen3.5-397B-A17B","n":"Qwen3.5-397B-A17B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.6,"co":3.6,"cr":0.06,"cw":0.75,"ctx":262144,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-15","lu":"2026-05-07"},{"p":"nebius","id":"Qwen/Qwen3.5-397B-A17B-fast","n":"Qwen3.5-397B-A17B-fast","d":"Legacy model retained for compatibility with older integrations","ci":0.6,"co":3.6,"cr":0.06,"cw":0.75,"ctx":8000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-15","lu":"2026-05-07"},{"p":"nebius","id":"Qwen/Qwen3-30B-A3B-Instruct-2507","n":"Qwen3-30B-A3B-Instruct-2507","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.3,"cr":0.01,"cw":0.125,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2026-01-28","lu":"2026-02-04"},{"p":"nebius","id":"Qwen/Qwen3-Next-80B-A3B-Thinking","n":"Qwen3-Next-80B-A3B-Thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.15,"co":1.2,"cr":0.015,"cw":0.18,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-01-28","lu":"2026-02-04"},{"p":"nebius","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.2,"co":0.6,"ctx":262144,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2025-07","rd":"2025-07-25","lu":"2025-10-04"},{"p":"nebius","id":"Qwen/Qwen3-32B","n":"Qwen3-32B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.3,"cr":0.01,"cw":0.125,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2026-01-28","lu":"2026-02-04"},{"p":"nebius","id":"Qwen/Qwen3-235B-A22B-Thinking-2507-fast","n":"Qwen3-235B-A22B-Thinking-2507-fast","d":"Legacy model retained for compatibility with older integrations","ci":0.5,"co":2,"cr":0.05,"cw":0.625,"ctx":8000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-25","lu":"2026-05-07"},{"p":"nebius","id":"openai/gpt-oss-120b-fast","n":"gpt-oss-120b-fast","d":"Legacy model retained for compatibility with older integrations","ci":0.1,"co":0.5,"cr":0.01,"cw":0.125,"ctx":8000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-06-10","lu":"2026-05-07"},{"p":"nebius","id":"openai/gpt-oss-120b","n":"gpt-oss-120b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"cr":0.015,"cw":0.18,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2026-01-10","lu":"2026-02-04"},{"p":"nebius","id":"NousResearch/Hermes-4-405B","n":"Hermes-4-405B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1,"co":3,"cr":0.1,"cw":1.25,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-11","rd":"2026-01-30","lu":"2026-02-04"},{"p":"nebius","id":"NousResearch/Hermes-4-70B","n":"Hermes-4-70B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.13,"co":0.4,"cr":0.013,"cw":0.16,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-11","rd":"2026-01-30","lu":"2026-02-04"},{"p":"nebius","id":"nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B","n":"Nemotron-3-Nano-30B-A3B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.06,"co":0.24,"cr":0.006,"cw":0.075,"ctx":32000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-05","rd":"2025-08-10","lu":"2026-02-04"},{"p":"nebius","id":"nvidia/Nemotron-3-Nano-Omni","n":"Nemotron-3-Nano-Omni","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0.06,"co":0.24,"cr":0.006,"cw":0.075,"ctx":65536,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-01-20","lu":"2026-05-07"},{"p":"nebius","id":"nvidia/Llama-3_1-Nemotron-Ultra-253B-v1","n":"Llama-3.1-Nemotron-Ultra-253B-v1","f":"nemotron","d":"Flagship Nemotron model for high-throughput reasoning and complex agents","ci":0.6,"co":1.8,"cr":0.06,"cw":0.75,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-15","lu":"2026-02-04"},{"p":"nebius","id":"nvidia/nemotron-3-super-120b-a12b","n":"Nemotron-3-Super-120B-A12B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.3,"co":0.9,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-02","rd":"2026-03-11","lu":"2026-03-12"},{"p":"nebius","id":"zai-org/GLM-5","n":"GLM-5","d":"Legacy model retained for compatibility with older integrations","ci":1,"co":3.2,"cr":0.1,"cw":1,"ctx":200000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2026-01","rd":"2026-03-01","lu":"2026-03-10"},{"p":"nebius","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"ctx":432000,"out":432000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"nebius","id":"deepseek-ai/DeepSeek-V3.2-fast","n":"DeepSeek-V3.2-fast","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2,"cr":0.04,"cw":0.5,"ctx":8000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-01-27","lu":"2026-05-07"},{"p":"nebius","id":"deepseek-ai/DeepSeek-V4-Pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.75,"co":3.5,"cr":0.15,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"nebius","id":"deepseek-ai/DeepSeek-V3.2","n":"DeepSeek-V3.2","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":0.45,"cr":0.03,"cw":0.375,"ctx":163000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-11","rd":"2026-01-20","lu":"2026-02-04"},{"p":"nebius","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax-M2.5","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":196608,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-01-20","lu":"2026-05-07"},{"p":"nebius","id":"MiniMaxAI/MiniMax-M2.5-fast","n":"MiniMax-M2.5-fast","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":8000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-01-20","lu":"2026-05-07"},{"p":"nebius","id":"PrimeIntellect/INTELLECT-3","n":"INTELLECT-3","d":"Legacy model retained for compatibility with older integrations","ci":0.2,"co":1.1,"cr":0.02,"cw":0.25,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-10","rd":"2026-01-25","lu":"2026-02-04"},{"p":"auriko","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"auriko","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"auriko","id":"grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"auriko","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"auriko","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"auriko","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"auriko","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"auriko","id":"minimax-m2-7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"auriko","id":"minimax-m2-7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"auriko","id":"qwen-3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.5,"co":3,"cr":0.1,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"auriko","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.5,"co":2.8,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"auriko","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"auriko","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"auriko","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"auriko","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"stepfun-ai","id":"step-2-16k","n":"Step 2 (16K)","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":5.21,"co":16.44,"cr":1.04,"ctx":16384,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2025-01-01","lu":"2026-02-13"},{"p":"stepfun-ai","id":"step-tts-2","n":"Step TTS 2","f":"step","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2026-03-01","lu":"2026-07-02"},{"p":"stepfun-ai","id":"step-3.5-flash","n":"Step 3.5 Flash","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.1,"co":0.3,"cr":0.02,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-06-15"},{"p":"stepfun-ai","id":"stepaudio-2.5-asr","n":"StepAudio 2.5 ASR","f":"step","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ctx":0,"out":0,"mi":"a","mo":"t","rd":"2026-04-24","lu":"2026-07-02"},{"p":"stepfun-ai","id":"stepaudio-2.5-tts","n":"StepAudio 2.5 TTS","f":"step","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2026-04-16","lu":"2026-07-02"},{"p":"stepfun-ai","id":"step-3.5-flash-2603","n":"Step 3.5 Flash 2603","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"cr":0.02,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-02","lu":"2026-04-02"},{"p":"stepfun-ai","id":"step-3.7-flash","n":"Step 3.7 Flash","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0.2,"co":1.15,"cr":0.04,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01-01","rd":"2026-05-29","lu":"2026-06-29"},{"p":"stepfun-ai","id":"step-1-32k","n":"Step 1 (32K)","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":2.05,"co":9.59,"cr":0.41,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2025-01-01","lu":"2026-02-13"},{"p":"vivgrid","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"vivgrid","id":"gpt-5.4-nano","n":"GPT-5.4 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"vivgrid","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"vivgrid","id":"gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"vivgrid","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-24","lu":"2026-02-24"},{"p":"vivgrid","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"vivgrid","id":"gpt-5.4-mini","n":"GPT-5.4 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"vivgrid","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"vivgrid","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.03,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"vivgrid","id":"deepseek-v3.2","n":"DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":0.42,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"vivgrid","id":"gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":1,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"vivgrid","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-14","lu":"2026-01-14"},{"p":"vivgrid","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"tinfoil","id":"kimi-k2-6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":1.5,"co":5.25,"ctx":256000,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"tinfoil","id":"llama3-3-70b","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":1.75,"co":2.75,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"tinfoil","id":"gpt-oss-safeguard-120b","n":"gpt-oss-safeguard-120b","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.15,"co":0.6,"ctx":131000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-10-29","lu":"2025-10-29"},{"p":"tinfoil","id":"nomic-embed-text","n":"Nomic Embed Text v1.5","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.05,"co":0,"ctx":8192,"out":768,"mi":"t","mo":"t","ow":1,"rd":"2024-02","lu":"2024-02"},{"p":"tinfoil","id":"gpt-oss-120b","n":"gpt-oss-120b","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.15,"co":0.6,"ctx":131000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-08-05","lu":"2025-08-05"},{"p":"tinfoil","id":"glm-5-2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.5,"co":5.25,"ctx":384000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"tinfoil","id":"gemma4-31b","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.4,"co":1,"ctx":256000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"mistral","id":"codestral-latest","n":"Codestral (latest)","f":"codestral","d":"Mistral code model for completions, refactors, and developer IDE workflows","ci":0.3,"co":0.9,"ctx":256000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-05-29","lu":"2025-01-04"},{"p":"mistral","id":"mistral-large-latest","n":"Mistral Large (latest)","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0.5,"co":1.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"mistral","id":"open-mistral-7b","n":"Mistral 7B","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.25,"co":0.25,"ctx":8000,"out":8000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2023-09-27","lu":"2023-09-27"},{"p":"mistral","id":"devstral-small-2507","n":"Devstral Small","f":"devstral","d":"Legacy model retained for compatibility with older integrations","ci":0.1,"co":0.3,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-05","rd":"2025-07-10","lu":"2025-07-10"},{"p":"mistral","id":"ministral-3b-latest","n":"Ministral 3B (latest)","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.04,"co":0.04,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-10-01","lu":"2024-10-04"},{"p":"mistral","id":"pixtral-large-latest","n":"Pixtral Large (latest)","f":"pixtral","d":"Mistral's larger vision model for document-heavy image understanding and chat","ci":2,"co":6,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2024-11-04"},{"p":"mistral","id":"mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-07-01","lu":"2024-07-01"},{"p":"mistral","id":"mistral-embed","n":"Mistral Embed","f":"mistral-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":8000,"out":3072,"mi":"t","mo":"t","rd":"2023-12-11","lu":"2023-12-11"},{"p":"mistral","id":"mistral-small-2506","n":"Mistral Small 3.2","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-06-20","lu":"2025-06-20"},{"p":"mistral","id":"ministral-8b-latest","n":"Ministral 8B (latest)","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-10-01","lu":"2024-10-04"},{"p":"mistral","id":"open-mixtral-8x22b","n":"Mixtral 8x22B","f":"mixtral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":2,"co":6,"ctx":64000,"out":64000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-04-17","lu":"2024-04-17"},{"p":"mistral","id":"mistral-medium-latest","n":"Mistral Medium (latest)","f":"mistral-medium","d":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","ci":1.5,"co":7.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"mistral","id":"devstral-small-2505","n":"Devstral Small 2505","f":"devstral","d":"Legacy model retained for compatibility with older integrations","ci":0.1,"co":0.3,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-05","rd":"2025-05-07","lu":"2025-05-07"},{"p":"mistral","id":"magistral-small","n":"Magistral Small","f":"magistral-small","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":0.5,"co":1.5,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-03-17","lu":"2025-03-17"},{"p":"mistral","id":"mistral-medium-2604","n":"Mistral Medium 3.5","f":"mistral-medium","d":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","ci":1.5,"co":7.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"mistral","id":"mistral-small-latest","n":"Mistral Small (latest)","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.15,"co":0.6,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-03-16","lu":"2026-03-16"},{"p":"mistral","id":"open-mixtral-8x7b","n":"Mixtral 8x7B","f":"mixtral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.7,"co":0.7,"ctx":32000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-01","rd":"2023-12-11","lu":"2023-12-11"},{"p":"mistral","id":"devstral-latest","n":"Devstral 2","f":"devstral","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"mistral","id":"mistral-small-2603","n":"Mistral Small 4","f":"mistral-small","d":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","ci":0.15,"co":0.6,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-03-16","lu":"2026-03-16"},{"p":"mistral","id":"mistral-medium-2505","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":131072,"out":131072,"mi":"ti","mo":"t","tc":1,"kn":"2025-05","rd":"2025-05-07","lu":"2025-05-07"},{"p":"mistral","id":"mistral-large-2411","n":"Mistral Large 2.1","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-18","lu":"2024-11-18"},{"p":"mistral","id":"mistral-medium-2508","n":"Mistral Medium 3.1","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"kn":"2025-05","rd":"2025-08-12","lu":"2025-08-12"},{"p":"mistral","id":"open-mistral-nemo","n":"Open Mistral Nemo","f":"mistral-nemo","d":"Legacy model retained for compatibility with older integrations","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-07-01","lu":"2024-07-01"},{"p":"mistral","id":"magistral-medium-latest","n":"Magistral Medium (latest)","f":"magistral-medium","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":2,"co":5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2025-03-17","lu":"2025-03-20"},{"p":"mistral","id":"devstral-medium-latest","n":"Devstral 2 (latest)","f":"devstral","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-02","lu":"2025-12-02"},{"p":"mistral","id":"devstral-2512","n":"Devstral 2","f":"devstral","d":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"mistral","id":"labs-devstral-small-2512","n":"Devstral Small 2","f":"devstral","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"ctx":256000,"out":256000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"mistral","id":"pixtral-12b","n":"Pixtral 12B","f":"pixtral","d":"Mistral vision-language model for image understanding and multimodal chat","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-09","rd":"2024-09-01","lu":"2024-09-01"},{"p":"mistral","id":"mistral-large-2512","n":"Mistral Large 3","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ci":0.5,"co":1.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"mistral","id":"devstral-medium-2507","n":"Devstral Medium","f":"devstral","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-05","rd":"2025-07-10","lu":"2025-07-10"},{"p":"cloudflare-workers-ai","id":"@cf/ibm-granite/granite-4.0-h-micro","n":"Granite 4.0 H Micro","f":"granite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.017,"co":0.112,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-10-07","lu":"2025-10-07"},{"p":"cloudflare-workers-ai","id":"@cf/moonshotai/kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"cloudflare-workers-ai","id":"@cf/moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"cloudflare-workers-ai","id":"@cf/google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.1,"co":0.3,"ctx":256000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"cloudflare-workers-ai","id":"@cf/openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.35,"co":0.75,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"cloudflare-workers-ai","id":"@cf/openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.2,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"cloudflare-workers-ai","id":"@cf/mistralai/mistral-small-3.1-24b-instruct","n":"Mistral Small 3.1 24B Instruct","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.351,"co":0.555,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-03-18","lu":"2025-03-18"},{"p":"cloudflare-workers-ai","id":"@cf/nvidia/nemotron-3-120b-a12b","n":"Nemotron 3 Super 120B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.5,"co":1.5,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"cloudflare-workers-ai","id":"@cf/zai-org/glm-5.2","n":"Glm 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"cloudflare-workers-ai","id":"@cf/zai-org/glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.0605,"co":0.4,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"cloudflare-workers-ai","id":"@cf/deepseek-ai/deepseek-r1-distill-qwen-32b","n":"Deepseek R1 Distill Qwen 32B","f":"deepseek-thinking","d":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","ci":0.497,"co":4.881,"ctx":80000,"out":80000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"cloudflare-workers-ai","id":"@cf/qwen/qwen3-30b-a3b-fp8","n":"Qwen3 30B A3b fp8","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.0509,"co":0.335,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-30","lu":"2025-04-30"},{"p":"cloudflare-workers-ai","id":"@cf/qwen/qwen2.5-coder-32b-instruct","n":"Qwen2.5 Coder 32B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.66,"co":1,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2025-02-27","lu":"2025-02-27"},{"p":"cloudflare-workers-ai","id":"@cf/qwen/qwq-32b","n":"Qwq 32B","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.66,"co":1,"ctx":24000,"out":24000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-03-05","lu":"2025-03-05"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-3.2-1b-instruct","n":"Llama 3.2 1B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.027,"co":0.201,"ctx":60000,"out":60000,"mi":"t","mo":"t","ow":1,"rd":"2024-09-25","lu":"2024-09-25"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-3.2-11b-vision-instruct","n":"Llama 3.2 11B Vision Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.0485,"co":0.676,"ctx":128000,"out":128000,"mi":"ti","mo":"t","ow":1,"rd":"2024-09-25","lu":"2024-09-25"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout 17B 16E Instruct","f":"llama","d":"Open Llama with long-context vision for efficient multimodal agents","ci":0.27,"co":0.85,"ctx":131000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-guard-3-8b","n":"Llama Guard 3 8B","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.484,"co":0.03,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"rd":"2025-01-22","lu":"2025-01-22"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-3.3-70b-instruct-fp8-fast","n":"Llama 3.3 70B Instruct fp8 Fast","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":0.293,"co":2.253,"ctx":24000,"out":24000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-3.1-8b-instruct-fp8","n":"Llama 3.1 8B Instruct fp8","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.152,"co":0.287,"ctx":32000,"out":32000,"mi":"t","mo":"t","ow":1,"rd":"2024-07-25","lu":"2024-07-25"},{"p":"cloudflare-workers-ai","id":"@cf/meta/llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.0509,"co":0.335,"ctx":80000,"out":80000,"mi":"t","mo":"t","ow":1,"rd":"2024-09-25","lu":"2024-09-25"},{"p":"cloudflare-workers-ai","id":"@cf/aisingapore/gemma-sea-lion-v4-27b-it","n":"Gemma Sea Lion V4 27B It","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.351,"co":0.555,"ctx":128000,"out":128000,"mi":"t","mo":"t","ow":1,"rd":"2025-09-23","lu":"2025-09-23"},{"p":"bailing","id":"Ring-1T","n":"Ring-1T","f":"ring","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.57,"co":2.29,"ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-06","rd":"2025-10","lu":"2025-10"},{"p":"bailing","id":"Ling-1T","n":"Ling-1T","f":"ling","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.57,"co":2.29,"ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06","rd":"2025-10","lu":"2025-10"},{"p":"anyapi","id":"xai/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"anyapi","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"anyapi","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"anyapi","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"anyapi","id":"google/gemini-3-pro-preview","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"anyapi","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"anyapi","id":"openai/o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"anyapi","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"anyapi","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"anyapi","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"anyapi","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"anyapi","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"anyapi","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"anyapi","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"anyapi","id":"openai/gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"anyapi","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"anyapi","id":"mistralai/devstral-2512","n":"Devstral 2","f":"devstral","d":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"anyapi","id":"mistralai/mistral-large-2512","n":"Mistral Large 3","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"anyapi","id":"anthropic/claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"anyapi","id":"anthropic/claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"anyapi","id":"anthropic/claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"anyapi","id":"anthropic/claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"anyapi","id":"anthropic/claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"anyapi","id":"cohere/command-r-plus-08-2024","n":"Command R+","f":"command-r","d":"Cohere's RAG workhorse for long-context enterprise search and tool use","ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"anyapi","id":"perplexity/sonar-reasoning-pro","n":"Sonar Reasoning Pro","f":"sonar-reasoning","d":"Web-grounded Sonar for multi-step research questions that need cited reasoning","ctx":128000,"out":4096,"mi":"ti","mo":"t","rs":1,"kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"anyapi","id":"perplexity/sonar-pro","n":"Sonar Pro","f":"sonar-pro","d":"Deeper Sonar search model with broader retrieval and stronger synthesis","ctx":200000,"out":8192,"mi":"ti","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"anyapi","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"anyapi","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"anyapi","id":"deepseek/deepseek-r1","n":"DeepSeek Reasoner","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"anyapi","id":"deepseek/deepseek-chat","n":"DeepSeek Chat","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":1000000,"out":384000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"google","id":"gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"google","id":"gemini-2.5-flash-preview-tts","n":"Gemini 2.5 Flash Preview TTS","f":"gemini-flash","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0.5,"co":10,"ctx":8192,"out":16384,"mi":"t","mo":"a","kn":"2025-01","rd":"2025-05-01","lu":"2025-05-01"},{"p":"google","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"google","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"google","id":"gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"google","id":"gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"google","id":"gemini-2.0-flash","n":"Gemini 2.0 Flash","f":"gemini-flash","d":"Earlier Gemini Flash workhorse for responsive multimodal apps and tool use","ci":0.1,"co":0.4,"cr":0.025,"ctx":1048576,"out":8192,"mi":"tiavp","mo":"t","tc":1,"kn":"2024-06","rd":"2024-12-11","lu":"2024-12-11"},{"p":"google","id":"gemini-embedding-001","n":"Gemini Embedding 001","f":"gemini","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.15,"co":0,"ctx":2048,"out":1,"mi":"t","mo":"t","kn":"2025-05","rd":"2025-05-20","lu":"2025-05-20"},{"p":"google","id":"gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"google","id":"gemini-flash-lite-latest","n":"Gemini Flash-Lite Latest","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"google","id":"gemini-3-pro-image-preview","n":"Nano Banana Pro","f":"gemini-pro","d":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","ci":2,"co":120,"ctx":131072,"out":32768,"mi":"ti","mo":"ti","rs":1,"kn":"2025-01","rd":"2025-11-20","lu":"2025-11-20"},{"p":"google","id":"gemini-2.5-flash-image","n":"Nano Banana","f":"gemini-flash","d":"Nano Banana image model for fast generation, edits, and character-consistent assets","ci":0.3,"co":30,"cr":0.075,"ctx":32768,"out":32768,"mi":"ti","mo":"ti","rs":1,"kn":"2025-06","rd":"2025-08-26","lu":"2025-08-26"},{"p":"google","id":"gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"google","id":"gemini-3.1-flash-image-preview","n":"Nano Banana 2","f":"gemini-flash","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":60,"ctx":65536,"out":65536,"mi":"tip","mo":"ti","rs":1,"kn":"2025-01","rd":"2026-02-26","lu":"2026-02-26"},{"p":"google","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"google","id":"gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"google","id":"gemini-3-pro-preview","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"google","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"google","id":"gemini-2.5-pro-preview-tts","n":"Gemini 2.5 Pro Preview TTS","f":"gemini-flash","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1,"co":20,"ctx":8192,"out":16384,"mi":"t","mo":"a","kn":"2025-01","rd":"2025-05-01","lu":"2025-05-01"},{"p":"google","id":"gemini-flash-latest","n":"Gemini Flash Latest","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"google","id":"gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Legacy model retained for compatibility with older integrations","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"google","id":"gemini-2.0-flash-lite","n":"Gemini 2.0 Flash-Lite","f":"gemini-flash-lite","d":"Legacy model retained for compatibility with older integrations","ci":0.075,"co":0.3,"ctx":1048576,"out":8192,"mi":"tiavp","mo":"t","tc":1,"kn":"2024-06","rd":"2024-12-11","lu":"2024-12-11"},{"p":"opencode-go","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode-go","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax-m2.5","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":1.2,"cr":0.03,"ctx":204800,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"opencode-go","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen3.7-plus","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.4,"co":1.6,"cr":0.04,"cw":0.5,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-06-02","lu":"2026-06-02"},{"p":"opencode-go","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen3.7-max","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":2.5,"co":7.5,"cr":0.5,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"opencode-go","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"opencode-go","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cr":0.26,"ctx":202752,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-07","lu":"2026-04-07"},{"p":"opencode-go","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.74,"co":3.48,"cr":0.0145,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode-go","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"opencode-go","id":"minimax-m3","n":"MiniMax-M3","f":"minimax-m3","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":1000000,"out":131072,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-05-31","lu":"2026-05-31"},{"p":"opencode-go","id":"qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen3.5","d":"Legacy model retained for compatibility with older integrations","ci":0.2,"co":1.2,"cr":0.02,"cw":0.25,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"opencode-go","id":"minimax-m2.7","n":"MiniMax-M2.7","f":"minimax-m2.7","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-03-18","lu":"2026-03-18"},{"p":"opencode-go","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Legacy model retained for compatibility with older integrations","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2026-01-27","lu":"2026-01-27"},{"p":"opencode-go","id":"mimo-v2.5","n":"MiMo V2.5","f":"mimo-v2.5","d":"MiMo omni model for text, image, video, audio, and agents","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":128000,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"opencode-go","id":"mimo-v2-omni","n":"MiMo V2 Omni","f":"mimo-v2-omni","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2,"cr":0.08,"ctx":262144,"out":128000,"mi":"tiap","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"opencode-go","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2026-04-21","lu":"2026-04-21"},{"p":"opencode-go","id":"mimo-v2-pro","n":"MiMo V2 Pro","f":"mimo-v2-pro","d":"Legacy model retained for compatibility with older integrations","ci":1,"co":3,"cr":0.2,"ctx":1048576,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"opencode-go","id":"mimo-v2.5-pro","n":"MiMo V2.5 Pro","f":"mimo-v2.5-pro","d":"MiMo pro model for strong multimodal reasoning and agent execution","ci":1.74,"co":3.48,"cr":0.0145,"ctx":1048576,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"opencode-go","id":"glm-5","n":"GLM-5","f":"glm","d":"Legacy model retained for compatibility with older integrations","ci":1,"co":3.2,"cr":0.2,"ctx":202752,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-11","lu":"2026-02-11"},{"p":"opencode-go","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen3.6","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"digitalocean","id":"anthropic-claude-haiku-4.5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"digitalocean","id":"openai-gpt-image-1","n":"GPT Image 1","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":40,"cr":1.25,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-04-24","lu":"2025-04-24"},{"p":"digitalocean","id":"e5-large-v2","n":"E5 Large v2","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.02,"co":0,"ctx":512,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2023-05-19","lu":"2026-04-30"},{"p":"digitalocean","id":"bge-m3","n":"BGE M3","f":"bge","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.02,"co":0,"ctx":8192,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2024-01-30","lu":"2026-04-30"},{"p":"digitalocean","id":"mistral-3-14B","n":"Ministral 3 14B Instruct","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"ctx":262144,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-15","lu":"2026-04-30"},{"p":"digitalocean","id":"nemotron-3-ultra-550b","n":"Nemotron 3 Ultra","f":"nemotron","d":"Flagship Nemotron model for high-throughput reasoning and complex agents","ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2026-06-04","lu":"2026-06-12"},{"p":"digitalocean","id":"minimax-m2.5","n":"MiniMax M2.5","f":"minimax-m2.5","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2026-02-12","lu":"2026-04-16"},{"p":"digitalocean","id":"openai-gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"digitalocean","id":"deepseek-v3","n":"DeepSeek V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":163840,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-12-26","lu":"2025-03-24"},{"p":"digitalocean","id":"openai-gpt-image-2","n":"GPT Image 2","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-04-24","lu":"2025-04-24"},{"p":"digitalocean","id":"openai-gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"digitalocean","id":"deepseek-r1-distill-llama-70b","n":"DeepSeek R1 Distill Llama 70B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.99,"co":0.99,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-01-30","lu":"2025-01-30"},{"p":"digitalocean","id":"qwen3-embedding-0.6b","n":"Qwen3 Embedding 0.6B","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.04,"co":0,"ctx":8000,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2025-06-03","lu":"2026-04-16"},{"p":"digitalocean","id":"gemma-4-31B-it","n":"Gemma 4 31B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.18,"co":0.5,"ctx":256000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-30"},{"p":"digitalocean","id":"llama-4-maverick","n":"Llama 4 Maverick 17B 128E Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.25,"co":0.87,"ctx":1000000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2026-04-30"},{"p":"digitalocean","id":"anthropic-claude-3.7-sonnet","n":"Claude 3.7 Sonnet","f":"claude-sonnet","d":"Legacy model retained for compatibility with older integrations","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-11","rd":"2025-02-24","lu":"2025-02-24"},{"p":"digitalocean","id":"openai-gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"digitalocean","id":"anthropic-claude-opus-4.7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"digitalocean","id":"deepseek-4-flash","n":"Deepseek V4 Flash","f":"deepseek","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ctx":262144,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2026-05-27","lu":"2026-05-29"},{"p":"digitalocean","id":"anthropic-claude-4.5-haiku","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":1,"cw":1.25,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"digitalocean","id":"anthropic-claude-4.6-sonnet","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"digitalocean","id":"anthropic-claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"digitalocean","id":"ministral-3-8b-instruct-2512","n":"Ministral 3 8B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-15","lu":"2025-12-15"},{"p":"digitalocean","id":"anthropic-claude-4.5-sonnet","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"digitalocean","id":"anthropic-claude-opus-4.6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"digitalocean","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.975,"co":4.3,"cr":0.26,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"digitalocean","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.74,"co":3.48,"ctx":1048576,"out":393216,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"digitalocean","id":"openai-gpt-oss-20b","n":"gpt-oss-20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.45,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-08-05","lu":"2026-04-16"},{"p":"digitalocean","id":"qwen-2.5-14b-instruct","n":"Qwen 2.5 14B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-09","rd":"2024-09-19","lu":"2024-09-19"},{"p":"digitalocean","id":"anthropic-claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-29"},{"p":"digitalocean","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.05,"co":4.4,"cr":0.21,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"digitalocean","id":"anthropic-claude-opus-4.5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"digitalocean","id":"anthropic-claude-4.1-opus","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"digitalocean","id":"llama3-8b-instruct","n":"Llama 3.1 Instruct (8B)","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.198,"co":0.198,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"digitalocean","id":"stable-diffusion-3.5-large","n":"Stable Diffusion 3.5 Large","f":"stable-diffusion","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.08,"co":0,"ctx":256,"out":1,"mi":"t","mo":"i","ow":1,"rd":"2024-10-22","lu":"2026-04-30"},{"p":"digitalocean","id":"openai-gpt-5.4-pro","n":"GPT-5.4 pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"digitalocean","id":"all-mini-lm-l6-v2","n":"All-MiniLM-L6-v2","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.009,"co":0,"ctx":256,"out":384,"mi":"t","mo":"t","ow":1,"rd":"2021-08-30","lu":"2026-04-16"},{"p":"digitalocean","id":"openai-gpt-image-1.5","n":"GPT Image 1.5","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":10,"cr":1,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-11-25","lu":"2025-11-25"},{"p":"digitalocean","id":"openai-gpt-4o","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"digitalocean","id":"anthropic-claude-opus-4","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"digitalocean","id":"bge-reranker-v2-m3","n":"BGE Reranker v2 M3","f":"bge","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0.01,"co":0,"ctx":8192,"out":1,"mi":"t","mo":"t","ow":1,"rd":"2024-03-12","lu":"2026-04-30"},{"p":"digitalocean","id":"openai-gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"digitalocean","id":"openai-o3-mini","n":"o3-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"digitalocean","id":"multi-qa-mpnet-base-dot-v1","n":"Multi-QA-mpnet-base-dot-v1","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.009,"co":0,"ctx":512,"out":768,"mi":"t","mo":"t","ow":1,"rd":"2021-08-30","lu":"2026-04-16"},{"p":"digitalocean","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.5,"co":2.7,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-04-16"},{"p":"digitalocean","id":"llama3.3-70b-instruct","n":"Llama 3.3 Instruct 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.65,"co":0.65,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"digitalocean","id":"gte-large-en-v1.5","n":"GTE Large (v1.5)","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.09,"co":0,"ctx":8192,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2024-03-27","lu":"2026-04-16"},{"p":"digitalocean","id":"openai-gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"digitalocean","id":"openai-gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-30"},{"p":"digitalocean","id":"nemotron-3-nano-30b","n":"Nemotron 3 Nano 30B A3B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-14","lu":"2025-04-14"},{"p":"digitalocean","id":"openai-gpt-oss-120b","n":"gpt-oss-120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.1,"co":0.7,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-08-05","lu":"2026-04-16"},{"p":"digitalocean","id":"openai-gpt-5-nano","n":"GPT-5 nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"digitalocean","id":"wan2-2-t2v-a14b","n":"Wan2.2-T2V-A14B","d":"Video model for prompt-guided generation, editing, and motion workflows","ci":0.6,"co":0,"ctx":100,"out":1,"mi":"t","mo":"v","ow":1,"rd":"2025-07-28","lu":"2026-04-30"},{"p":"digitalocean","id":"mistral-7b-instruct-v0.3","n":"Mistral 7B Instruct v0.3","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-05-22","lu":"2024-05-22"},{"p":"digitalocean","id":"qwen3-tts-voicedesign","n":"Qwen3 TTS VoiceDesign","f":"qwen","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":32768,"out":1,"mi":"t","mo":"a","ow":1,"rd":"2026-04-21","lu":"2026-04-30"},{"p":"digitalocean","id":"mistral-nemo-instruct-2407","n":"Mistral Nemo Instruct","f":"mistral","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-18","lu":"2024-07-18"},{"p":"digitalocean","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"digitalocean","id":"deepseek-3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.5,"co":1.6,"ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-12-02","lu":"2026-04-30"},{"p":"digitalocean","id":"openai-o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"digitalocean","id":"qwen3.5-397b-a17b","n":"Qwen 3.5 397B A17B","f":"qwen3.5","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.55,"co":3.5,"ctx":262144,"out":81920,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-04-30"},{"p":"digitalocean","id":"qwen3-coder-flash","n":"Qwen3 Coder Flash","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.45,"co":1.7,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2026-04-30"},{"p":"digitalocean","id":"openai-gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"digitalocean","id":"openai-gpt-4.1","n":"GPT-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"digitalocean","id":"anthropic-claude-3.5-haiku","n":"Claude 3.5 Haiku","f":"claude-haiku","d":"Legacy model retained for compatibility with older integrations","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-11-05","lu":"2024-11-05"},{"p":"digitalocean","id":"openai-gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"digitalocean","id":"arcee-trinity-large-thinking","n":"Trinity Large Thinking","f":"trinity","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":0.9,"cr":0.06,"ctx":256000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-16"},{"p":"digitalocean","id":"openai-o3","n":"o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"digitalocean","id":"anthropic-claude-3.5-sonnet","n":"Claude 3.5 Sonnet","f":"claude-sonnet","d":"Legacy model retained for compatibility with older integrations","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-06-20","lu":"2024-10-22"},{"p":"digitalocean","id":"openai-gpt-5-mini","n":"GPT-5 mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"digitalocean","id":"nvidia-nemotron-3-super-120b","n":"Nemotron-3-Super-120B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.3,"co":0.65,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-02","rd":"2026-03-11","lu":"2026-04-16"},{"p":"digitalocean","id":"glm-5","n":"GLM 5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"ctx":202752,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-04-16"},{"p":"digitalocean","id":"openai-gpt-5.2-pro","n":"GPT-5.2 pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"digitalocean","id":"anthropic-claude-3-opus","n":"Claude 3 Opus","f":"claude-opus","d":"Legacy model retained for compatibility with older integrations","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-08","rd":"2024-02-29","lu":"2024-02-29"},{"p":"digitalocean","id":"anthropic-claude-fable-5","n":"Anthropic Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-09","lu":"2026-06-12"},{"p":"digitalocean","id":"nemotron-3-nano-omni","n":"Nemotron Nano 3 Omni","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0.5,"co":0.9,"ctx":65536,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-30"},{"p":"digitalocean","id":"alibaba-qwen3-32b","n":"Qwen3-32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.25,"co":0.55,"ctx":131000,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-30","lu":"2026-04-16"},{"p":"digitalocean","id":"nemotron-nano-12b-v2-vl","n":"Nemotron Nano 12B v2 VL","f":"nemotron","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0.2,"co":0.6,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2025-12-01","lu":"2026-04-30"},{"p":"digitalocean","id":"openai-gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"digitalocean","id":"fal-ai/fast-sdxl","n":"Fast SDXL","f":"stable-diffusion","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","ow":1,"rd":"2023-07-26","lu":"2026-04-16"},{"p":"digitalocean","id":"fal-ai/elevenlabs/tts/multilingual-v2","n":"ElevenLabs Multilingual TTS v2","f":"elevenlabs","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2023-08-22","lu":"2026-04-16"},{"p":"digitalocean","id":"fal-ai/flux/schnell","n":"FLUX.1 [schnell]","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"t","mo":"i","ow":1,"rd":"2024-08-01","lu":"2026-04-16"},{"p":"digitalocean","id":"fal-ai/stable-audio-25/text-to-audio","n":"Stable Audio 2.5 (Text-to-Audio)","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2025-10-08","lu":"2026-04-16"},{"p":"subconscious","id":"subconscious/glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"subconscious","id":"subconscious/tim-qwen3.6-27b","n":"TIM-Qwen3.6 27B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.3,"co":3,"cr":0.15,"ctx":8192,"out":5000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"venice","id":"z-ai-glm-5-turbo","n":"GLM 5 Turbo","f":"glm","d":"Faster GLM-5 lane for coding agents that need lower latency","ci":1.2,"co":4,"cr":0.24,"ctx":200000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-15","lu":"2026-06-11"},{"p":"venice","id":"grok-4-20-multi-agent","n":"Grok 4.20 Multi-Agent","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.42,"co":2.83,"cr":0.23,"ctx":2000000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2026-03-12","lu":"2026-06-11"},{"p":"venice","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.138,"co":0.275,"cr":0.028,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-06-11"},{"p":"venice","id":"google-gemma-4-31b-it","n":"Google Gemma 4 31B Instruct","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.12,"co":0.36,"cr":0.09,"ctx":256000,"out":8192,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-03","lu":"2026-06-11"},{"p":"venice","id":"kimi-k2-6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.75,"co":3.5,"cr":0.16,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-20","lu":"2026-06-11"},{"p":"venice","id":"qwen3-235b-a22b-instruct-2507","n":"Qwen 3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":0.75,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-29","lu":"2026-06-11"},{"p":"venice","id":"nvidia-nemotron-cascade-2-30b-a3b","n":"Nemotron Cascade 2 30B A3B","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.14,"co":0.8,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-24","lu":"2026-06-11"},{"p":"venice","id":"claude-opus-4-7-fast","n":"Claude Opus 4.7 Fast","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":36,"co":180,"cr":3.6,"cw":45,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-05-14","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-55-pro","n":"GPT-5.5 Pro","f":"gpt","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":37.5,"co":225,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-24","lu":"2026-06-11"},{"p":"venice","id":"llama-3.3-70b","n":"Llama 3.3 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.7,"co":2.8,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-06","lu":"2026-06-11"},{"p":"venice","id":"qwen3-5-397b-a17b","n":"Qwen 3.5 397B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.75,"co":4.5,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-16","lu":"2026-06-11"},{"p":"venice","id":"claude-opus-4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":6,"co":30,"cr":0.6,"cw":7.5,"ctx":198000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-12-06","lu":"2026-06-11"},{"p":"venice","id":"zai-org-glm-5","n":"GLM 5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":1,"co":3.2,"cr":0.2,"ctx":198000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-06-11"},{"p":"venice","id":"qwen3-5-35b-a3b","n":"Qwen 3.5 35B A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3125,"co":1.25,"cr":0.15625,"ctx":256000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-25","lu":"2026-06-11"},{"p":"venice","id":"venice-uncensored-role-play","n":"Venice Role Play Uncensored","f":"venice","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.5,"co":2,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-02-20","lu":"2026-06-11"},{"p":"venice","id":"qwen-3-6-plus","n":"Qwen 3.6 Plus Uncensored","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.625,"co":3.75,"cr":0.0625,"cw":0.78,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-06","lu":"2026-06-11"},{"p":"venice","id":"zai-org-glm-4.6","n":"GLM 4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.43,"co":1.75,"cr":0.08,"ctx":198000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2024-04-01","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-4o-2024-11-20","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":3.125,"co":12.5,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2026-02-28","lu":"2026-06-11"},{"p":"venice","id":"grok-4-3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.42,"co":2.83,"cr":0.23,"ctx":1000000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-18","lu":"2026-06-11"},{"p":"venice","id":"qwen-3-7-plus","n":"Qwen 3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.5,"co":2,"cr":0.05,"cw":0.625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-11"},{"p":"venice","id":"minimax-m27","n":"MiniMax M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.375,"co":1.5,"cr":0.06875,"ctx":198000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-06-11"},{"p":"venice","id":"kimi-k2-7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.75,"co":3.5,"cr":0.16,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-13","lu":"2026-06-16"},{"p":"venice","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.65,"co":3.301,"cr":0.33,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-06-11"},{"p":"venice","id":"qwen3-235b-a22b-thinking-2507","n":"Qwen 3 235B A22B Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.45,"co":3.5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-29","lu":"2026-06-11"},{"p":"venice","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3.75,"co":18.75,"cr":0.375,"cw":4.69,"ctx":198000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-01-15","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-4o-mini-2024-07-18","n":"GPT-4o Mini","f":"gpt","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.1875,"co":0.75,"cr":0.09375,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2026-02-28","lu":"2026-06-11"},{"p":"venice","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":6,"co":30,"cr":0.6,"cw":7.5,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-06-11"},{"p":"venice","id":"claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-29","lu":"2026-07-01"},{"p":"venice","id":"openai-gpt-53-codex","n":"GPT-5.3 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":2.19,"co":17.5,"cr":0.219,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-24","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-54","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":3.13,"co":18.8,"cr":0.313,"ctx":1000000,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-06-11"},{"p":"venice","id":"google-gemma-3-27b-it","n":"Google Gemma 3 27B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.12,"co":0.2,"ctx":198000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-11-04","lu":"2026-06-11"},{"p":"venice","id":"claude-opus-4-8-fast","n":"Claude Opus 4.8 Fast","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":12,"co":60,"cr":1.2,"cw":15,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-06-11"},{"p":"venice","id":"gemma-4-uncensored","n":"Gemma 4 Uncensored","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.1625,"co":0.5,"ctx":256000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-04-13","lu":"2026-06-11"},{"p":"venice","id":"qwen3-next-80b","n":"Qwen 3 Next 80b","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.35,"co":1.9,"ctx":256000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-29","lu":"2026-06-11"},{"p":"venice","id":"mistral-small-2603","n":"Mistral Small 4","f":"mistral-small","d":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","ci":0.1875,"co":0.75,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-03-16","lu":"2026-06-11"},{"p":"venice","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":6,"co":30,"cr":0.6,"cw":7.5,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-06-11"},{"p":"venice","id":"zai-org-glm-4.7-flash","n":"GLM 4.7 Flash","f":"glm","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0.125,"co":0.5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-29","lu":"2026-06-11"},{"p":"venice","id":"google-gemma-4-26b-a4b-it","n":"Google Gemma 4 26B A4B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.13,"co":0.4,"cr":0.05,"ctx":256000,"out":8192,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-06-11"},{"p":"venice","id":"grok-4-20","n":"Grok 4.20","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.42,"co":2.83,"cr":0.23,"ctx":2000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-12","lu":"2026-06-11"},{"p":"venice","id":"hermes-3-llama-3.1-405b","n":"Hermes 3 Llama 3.1 405b","f":"hermes","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":1.1,"co":3,"ctx":128000,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2025-09-25","lu":"2026-06-11"},{"p":"venice","id":"aion-labs-aion-2-0","n":"Aion 2.0","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1,"co":2,"cr":0.25,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2026-03-24","lu":"2026-06-11"},{"p":"venice","id":"qwen3-vl-235b-a22b","n":"Qwen3 VL 235B","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.21,"co":1.9,"cr":0.1,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-01-16","lu":"2026-06-11"},{"p":"venice","id":"xiaomi-mimo-v2-5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.14,"co":0.28,"cr":0.05,"ctx":1000000,"out":65536,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-06-11","lu":"2026-06-11"},{"p":"venice","id":"venice-uncensored-1-2","n":"Venice Uncensored 1.2","f":"venice","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.2,"co":0.9,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-06-11"},{"p":"venice","id":"gemini-3-5-flash","n":"Gemini 3.5 Flash","f":"gemini","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.55,"co":9.45,"cr":0.155,"cw":0.086,"ctx":1000000,"out":65536,"mi":"tiav","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-22","lu":"2026-06-11"},{"p":"venice","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":12,"co":60,"cr":1.2,"cw":15,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-10","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-oss-120b","n":"OpenAI GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.07,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-06","lu":"2026-06-11"},{"p":"venice","id":"zai-org-glm-5-2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-16","lu":"2026-06-16"},{"p":"venice","id":"nvidia-nemotron-3-nano-30b-a3b","n":"NVIDIA Nemotron 3 Nano 30B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.075,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-06-11"},{"p":"venice","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":6,"co":30,"cr":0.6,"cw":7.5,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-06-11"},{"p":"venice","id":"mistral-small-3-2-24b-instruct","n":"Mistral Small 3.2 24B Instruct","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.09375,"co":0.25,"ctx":256000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01-15","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-54-pro","n":"GPT-5.4 Pro","f":"gpt","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":37.5,"co":225,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-06-11"},{"p":"venice","id":"minimax-m3-preview","n":"MiniMax M3 Preview","f":"minimax-m3","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":524288,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-12","lu":"2026-06-13"},{"p":"venice","id":"gemini-3-1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2.5,"co":15,"cr":0.5,"cw":0.5,"ctx":1000000,"out":32768,"mi":"tiav","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-06-11"},{"p":"venice","id":"qwen3-coder-480b-a35b-instruct-turbo","n":"Qwen 3 Coder 480B Turbo","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.35,"co":1.5,"cr":0.04,"ctx":256000,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-06-11"},{"p":"venice","id":"qwen3-6-27b","n":"Qwen 3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.325,"co":3.25,"ctx":256000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-52-codex","n":"GPT-5.2 Codex","f":"gpt","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":2.19,"co":17.5,"cr":0.219,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08","rd":"2025-01-15","lu":"2026-06-11"},{"p":"venice","id":"qwen-3-7-max","n":"Qwen 3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.7,"co":8.05,"cr":0.27,"cw":3.35,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-05-22","lu":"2026-06-11"},{"p":"venice","id":"olafangensan-glm-4.7-flash-heretic","n":"GLM 4.7 Flash Heretic","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"cr":0.035,"ctx":200000,"out":24000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-04","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-54-mini","n":"GPT-5.4 Mini","f":"gpt","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.9375,"co":5.625,"cr":0.09375,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-27","lu":"2026-06-11"},{"p":"venice","id":"grok-build-0-1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-06-11"},{"p":"venice","id":"zai-org-glm-4.7","n":"GLM 4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.55,"co":2.65,"cr":0.11,"ctx":198000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-24","lu":"2026-06-11"},{"p":"venice","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3.6,"co":18,"cr":0.36,"cw":4.5,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-06-11"},{"p":"venice","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.7,"co":3.75,"cr":0.07,"ctx":256000,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-19","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-55","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":6.25,"co":37.5,"cr":0.625,"ctx":1000000,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-06-11"},{"p":"venice","id":"qwen3-5-9b","n":"Qwen 3.5 9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.15,"ctx":256000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-05","lu":"2026-06-11"},{"p":"venice","id":"mercury-2","n":"Mercury 2","f":"mercury","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.3125,"co":0.9375,"cr":0.03125,"ctx":128000,"out":50000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-20","lu":"2026-06-11"},{"p":"venice","id":"zai-org-glm-5-1","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.54,"co":4.84,"cr":0.286,"ctx":200000,"out":24000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-06-11"},{"p":"venice","id":"kimi-k2-5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.56,"co":3.5,"cr":0.22,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2026-01-27","lu":"2026-06-11"},{"p":"venice","id":"z-ai-glm-5v-turbo","n":"GLM 5V Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":1.5,"co":5,"cr":0.3,"ctx":200000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-06-11"},{"p":"venice","id":"nvidia-nemotron-3-ultra-550b-a55b","n":"NVIDIA Nemotron 3 Ultra","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.625,"co":3.125,"cr":0.1875,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-11"},{"p":"venice","id":"openai-gpt-52","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":2.19,"co":17.5,"cr":0.219,"ctx":256000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-13","lu":"2026-06-11"},{"p":"venice","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.33,"co":0.48,"cr":0.16,"ctx":160000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-04","lu":"2026-06-11"},{"p":"venice","id":"minimax-m25","n":"MiniMax M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.27,"co":0.95,"cr":0.03,"ctx":198000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-06-11"},{"p":"venice","id":"llama-3.2-3b","n":"Llama 3.2 3B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.15,"co":0.6,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-10-03","lu":"2026-06-11"},{"p":"lmstudio","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"lmstudio","id":"qwen/qwen3-30b-a3b-2507","n":"Qwen3 30B A3B 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-30","lu":"2025-07-30"},{"p":"lmstudio","id":"qwen/qwen3-coder-30b","n":"Qwen3 Coder 30B","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"poolside","id":"poolside/laguna-xs.2","n":"Laguna XS.2","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-06-13"},{"p":"poolside","id":"poolside/laguna-m.1","n":"Laguna M.1","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-06-13"},{"p":"zenmux","id":"inclusionai/ling-1t","n":"Ling-1T","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.56,"co":2.24,"cr":0.11,"ctx":128000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-10-09","lu":"2025-10-09"},{"p":"zenmux","id":"inclusionai/ring-2.6-1t","n":"inclusionAI: Ring-2.6-1T","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.3,"co":2.5,"cr":0.06,"ctx":262000,"out":65000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12-31","rd":"2026-05-07","lu":"2026-05-14"},{"p":"zenmux","id":"inclusionai/ring-1t","n":"Ring-1T","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.56,"co":2.24,"cr":0.11,"ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-10-12","lu":"2025-10-12"},{"p":"zenmux","id":"moonshotai/kimi-k2.7-code-free","n":"Kimi K2.7 Code (Free)","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"cr":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"zenmux","id":"moonshotai/kimi-k2-thinking-turbo","n":"Kimi K2 Thinking Turbo","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":1.15,"co":8,"cr":0.15,"ctx":262000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-06","lu":"2025-11-06"},{"p":"zenmux","id":"moonshotai/kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"zenmux","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"cr":0.15,"ctx":262000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-06","lu":"2025-11-06"},{"p":"zenmux","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.58,"co":3.02,"cr":0.1,"ctx":262000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"zenmux","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262140,"out":262140,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01-01","rd":"2026-04-20","lu":"2026-04-20"},{"p":"zenmux","id":"moonshotai/kimi-k2-0905","n":"Kimi K2 0905","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"cr":0.15,"ctx":262000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-09-04","lu":"2025-09-04"},{"p":"zenmux","id":"baidu/ernie-5.0-thinking-preview","n":"ERNIE 5.0","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.84,"co":3.37,"ctx":128000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-01-22","lu":"2026-01-22"},{"p":"zenmux","id":"google/gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"zenmux","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.31,"cw":4.5,"ctx":1048000,"out":64000,"mi":"pitav","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"zenmux","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.07,"cw":1,"ctx":1048000,"out":64000,"mi":"pita","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"zenmux","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"zenmux","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.03,"cw":1,"ctx":1048000,"out":64000,"mi":"pita","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-07-22","lu":"2025-07-22"},{"p":"zenmux","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"cw":4.5,"ctx":1048000,"out":64000,"mi":"tipav","mo":"t","rs":1,"tc":1,"kn":"2026-02-19","rd":"2026-02-19","lu":"2026-02-19"},{"p":"zenmux","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash Preview","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"cw":1,"ctx":1048000,"out":64000,"mi":"tipa","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"zenmux","id":"google/gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"ctx":1050000,"out":65530,"mi":"tiav","mo":"t","tc":1,"rd":"2025-03-20","lu":"2025-03-20"},{"p":"zenmux","id":"x-ai/grok-code-fast-1","n":"Grok Code Fast 1","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":1.5,"cr":0.02,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-08-26","lu":"2025-08-26"},{"p":"zenmux","id":"x-ai/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"cw":0,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"zenmux","id":"x-ai/grok-4.1-fast-non-reasoning","n":"Grok 4.1 Fast Non Reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":64000,"mi":"ti","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-11-20","lu":"2025-11-20"},{"p":"zenmux","id":"x-ai/grok-4","n":"Grok 4","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"ctx":256000,"out":64000,"mi":"it","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-07-09","lu":"2025-07-09"},{"p":"zenmux","id":"x-ai/grok-4-fast","n":"Grok 4 Fast","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-09-19","lu":"2025-09-19"},{"p":"zenmux","id":"x-ai/grok-4.2-fast-non-reasoning","n":"Grok 4.2 Fast Non Reasoning","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":3,"co":9,"ctx":2000000,"out":30000,"mi":"tiv","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"x-ai/grok-4.2-fast","n":"Grok 4.2 Fast","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":3,"co":9,"ctx":2000000,"out":30000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"x-ai/grok-4.1-fast","n":"Grok 4.1 Fast","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-20","lu":"2025-11-20"},{"p":"zenmux","id":"x-ai/grok-build-0.1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"zenmux","id":"z-ai/glm-4.7","n":"GLM 4.7","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.28,"co":1.14,"cr":0.06,"ctx":200000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-23","lu":"2025-12-23"},{"p":"zenmux","id":"z-ai/glm-4.5","n":"GLM 4.5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.35,"co":1.54,"cr":0.07,"ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-07-25","lu":"2025-07-25"},{"p":"zenmux","id":"z-ai/glm-4.7-flashx","n":"GLM 4.7 FlashX","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.42,"cr":0.01,"ctx":200000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-01-19","lu":"2026-01-19"},{"p":"zenmux","id":"z-ai/glm-5.1","n":"GLM-5.1","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.8781,"co":3.5126,"cr":0.1903,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-03","lu":"2026-04-03"},{"p":"zenmux","id":"z-ai/glm-4.6","n":"GLM 4.6","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.35,"co":1.54,"cr":0.07,"ctx":200000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-09-30","lu":"2025-09-30"},{"p":"zenmux","id":"z-ai/glm-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.5,"cr":0.26,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"zenmux","id":"z-ai/glm-4.6v-flash-free","n":"GLM 4.6V Flash (Free)","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0,"co":0,"ctx":200000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-08","lu":"2025-12-08"},{"p":"zenmux","id":"z-ai/glm-4.7-flash-free","n":"GLM 4.7 Flash (Free)","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"ctx":200000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-01-19","lu":"2026-01-19"},{"p":"zenmux","id":"z-ai/glm-5.2-free","n":"GLM 5.2 (Free)","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"cr":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"zenmux","id":"z-ai/glm-4.6v","n":"GLM 4.6V","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.14,"co":0.42,"cr":0.03,"ctx":200000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-08","lu":"2025-12-08"},{"p":"zenmux","id":"z-ai/glm-5v-turbo","n":"GLM 5V Turbo","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.726,"co":3.1946,"cr":0.1743,"ctx":200000,"out":128000,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"zenmux","id":"z-ai/glm-4.6v-flash","n":"GLM 4.6V FlashX","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.02,"co":0.21,"cr":0.0043,"ctx":200000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-08","lu":"2025-12-08"},{"p":"zenmux","id":"z-ai/glm-4.5-air","n":"GLM 4.5 Air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.11,"co":0.56,"cr":0.02,"ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-07-25","lu":"2025-07-25"},{"p":"zenmux","id":"z-ai/glm-5","n":"GLM 5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.58,"co":2.6,"cr":0.14,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"zenmux","id":"z-ai/glm-5-turbo","n":"GLM 5 Turbo","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.88,"co":3.48,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.5-instant","n":"GPT-5.5 Instant","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":5,"co":30,"cr":0.5,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-05-05","lu":"2026-05-28"},{"p":"zenmux","id":"openai/gpt-5.2-pro","n":"GPT-5.2-Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"zenmux","id":"openai/gpt-5","n":"GPT-5","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.12,"ctx":400000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-08-07","lu":"2025-08-07"},{"p":"zenmux","id":"openai/gpt-5.1-chat","n":"GPT-5.1 Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.12,"ctx":128000,"out":64000,"mi":"pit","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-11-13","lu":"2025-11-13"},{"p":"zenmux","id":"openai/gpt-5.4-nano","n":"GPT-5.4 Nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"ctx":400000,"out":128000,"mi":"t","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.3-chat","n":"GPT-5.3 Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"ctx":128000,"out":16380,"mi":"t","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.1-codex","n":"GPT-5.1-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.12,"ctx":400000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-13","lu":"2025-11-13"},{"p":"zenmux","id":"openai/gpt-5.2","n":"GPT-5.2","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.17,"ctx":400000,"out":64000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-11","lu":"2025-12-11"},{"p":"zenmux","id":"openai/gpt-5.3-codex","n":"GPT-5.3 Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.1-codex-mini","n":"GPT-5.1-Codex-Mini","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.03,"ctx":400000,"out":64000,"mi":"it","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-13","lu":"2025-11-13"},{"p":"zenmux","id":"openai/gpt-5.4","n":"GPT-5.4","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":3.75,"co":18.75,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.4-mini","n":"GPT-5.4 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"ctx":400000,"out":128000,"mi":"t","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.4-pro","n":"GPT-5.4 Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":45,"co":225,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"openai/gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"zenmux","id":"openai/gpt-5-codex","n":"GPT-5 Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.12,"ctx":400000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-09-23","lu":"2025-09-23"},{"p":"zenmux","id":"openai/gpt-5.2-codex","n":"GPT-5.2-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.17,"ctx":400000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-01-15","lu":"2026-01-15"},{"p":"zenmux","id":"openai/gpt-5.1","n":"GPT-5.1","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.12,"ctx":400000,"out":64000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-13","lu":"2025-11-13"},{"p":"zenmux","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"zenmux","id":"xiaomi/mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.4,"co":2,"cr":0.08,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"zenmux","id":"xiaomi/mimo-v2-omni","n":"MiMo V2 Omni","f":"mimo","d":"MiMo omni model for text, image, video, audio, and agents","ci":0.4,"co":2,"cr":0.08,"ctx":265000,"out":265000,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"zenmux","id":"xiaomi/mimo-v2-flash","n":"MiMo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"cr":0.01,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12-01","rd":"2025-12-16","lu":"2026-02-04"},{"p":"zenmux","id":"xiaomi/mimo-v2-pro","n":"MiMo V2 Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":1,"co":3,"cr":0.2,"ctx":1000000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"zenmux","id":"xiaomi/mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":1,"co":3,"cr":0.2,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"zenmux","id":"sapiens-ai/agnes-1.5-pro","n":"Agnes 1.5 Pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.16,"co":0.8,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-21","lu":"2026-03-21"},{"p":"zenmux","id":"sapiens-ai/agnes-1.5-lite","n":"Agnes 1.5 Lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.12,"co":0.6,"ctx":256000,"out":256000,"mi":"ti","mo":"t","tc":1,"rd":"2026-03-26","lu":"2026-03-26"},{"p":"zenmux","id":"volcengine/doubao-seed-2.0-pro","n":"Doubao-Seed-2.0-pro","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.45,"co":2.24,"cr":0.09,"cw":0.0024,"ctx":256000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2026-02-14","rd":"2026-02-14","lu":"2026-02-14"},{"p":"zenmux","id":"volcengine/doubao-seed-code","n":"Doubao-Seed-Code","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.17,"co":1.12,"cr":0.03,"ctx":256000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-11","lu":"2025-11-11"},{"p":"zenmux","id":"volcengine/doubao-seed-1.8","n":"Doubao-Seed-1.8","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.11,"co":0.28,"cr":0.02,"cw":0.0024,"ctx":256000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-18","lu":"2025-12-18"},{"p":"zenmux","id":"volcengine/doubao-seed-2.0-mini","n":"Doubao-Seed-2.0-mini","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.03,"co":0.28,"cr":0.01,"cw":0.0024,"ctx":256000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2026-02-14","rd":"2026-02-14","lu":"2026-02-14"},{"p":"zenmux","id":"volcengine/doubao-seed-2.0-lite","n":"Doubao-Seed-2.0-lite","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.09,"co":0.51,"cr":0.02,"cw":0.0024,"ctx":256000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2026-02-14","rd":"2026-02-14","lu":"2026-02-14"},{"p":"zenmux","id":"volcengine/doubao-seed-2.0-code","n":"Doubao Seed 2.0 Code","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.9,"co":4.48,"ctx":256000,"out":32000,"mi":"t","mo":"t","tc":1,"kn":"2025-01-01","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"anthropic/claude-sonnet-5-free","n":"Claude Sonnet 5 (Free)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"zenmux","id":"anthropic/claude-3.5-haiku","n":"Claude 3.5 Haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":64000,"mi":"ti","mo":"t","tc":1,"kn":"2025-01-01","rd":"2024-11-04","lu":"2024-11-04"},{"p":"zenmux","id":"anthropic/claude-sonnet-4.5","n":"Claude Sonnet 4.5","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-09-29","lu":"2025-09-29"},{"p":"zenmux","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-05-22","lu":"2025-05-22"},{"p":"zenmux","id":"anthropic/claude-haiku-4.5","n":"Claude Haiku 4.5","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"it","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-10-15","lu":"2025-10-15"},{"p":"zenmux","id":"anthropic/claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":4,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"zenmux","id":"anthropic/claude-opus-4.7","n":"Claude Opus 4.7","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"zenmux","id":"anthropic/claude-3.7-sonnet","n":"Claude 3.7 Sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-02-24","lu":"2025-02-24"},{"p":"zenmux","id":"anthropic/claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"zenmux","id":"anthropic/claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"zenmux","id":"anthropic/claude-opus-4.1","n":"Claude Opus 4.1","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":64000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-08-05","lu":"2025-08-05"},{"p":"zenmux","id":"anthropic/claude-opus-4.5","n":"Claude Opus 4.5","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-11-24","lu":"2025-11-24"},{"p":"zenmux","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-18","lu":"2026-02-18"},{"p":"zenmux","id":"anthropic/claude-opus-4","n":"Claude Opus 4","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-05-22","lu":"2025-05-22"},{"p":"zenmux","id":"anthropic/claude-opus-4.6","n":"Claude Opus 4.6","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"it","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-06","lu":"2026-02-06"},{"p":"zenmux","id":"tencent/hy3-preview","n":"Hy3 preview","f":"Hy","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.172,"co":0.572,"cr":0.058,"cw":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"zenmux","id":"stepfun/step-3","n":"Step-3","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.21,"co":0.57,"ctx":65536,"out":64000,"mi":"it","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-07-31","lu":"2025-07-31"},{"p":"zenmux","id":"stepfun/step-3.7-flash","n":"Step 3.7 Flash","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0.2,"co":1.15,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01-01","rd":"2026-05-29","lu":"2026-05-29"},{"p":"zenmux","id":"stepfun/step-3.7-flash-free","n":"Step 3.7 Flash (Free)","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0,"co":0,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01-01","rd":"2026-05-29","lu":"2026-05-29"},{"p":"zenmux","id":"stepfun/step-3.5-flash","n":"Step 3.5 Flash","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"ctx":256000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2025-01-01","rd":"2026-02-02","lu":"2026-02-02"},{"p":"zenmux","id":"kuaishou/kat-coder-pro-v2","n":"KAT-Coder-Pro-V2","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":256000,"out":80000,"mi":"t","mo":"t","tc":1,"rd":"2026-03-30","lu":"2026-03-30"},{"p":"zenmux","id":"qwen/qwen3-coder-plus","n":"Qwen3-Coder-Plus","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":1000000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-07-23","lu":"2025-07-23"},{"p":"zenmux","id":"qwen/qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.4,"co":1.6,"cr":0.08,"cw":0.5,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"zenmux","id":"qwen/qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.5,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"zenmux","id":"qwen/qwen3-max","n":"Qwen3-Max-Thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":1.2,"co":6,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-01-23","lu":"2026-01-23"},{"p":"zenmux","id":"qwen/qwen3.5-flash","n":"Qwen3.5 Flash","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1,"co":0.4,"ctx":1020000,"out":1020000,"mi":"ti","mo":"t","tc":1,"kn":"2025-01-01","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"qwen/qwen3.5-plus","n":"Qwen3.5 Plus","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.8,"co":4.8,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"qwen/qwen3.6-plus","n":"Qwen3.6-Plus","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-30","lu":"2026-03-30"},{"p":"zenmux","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"zenmux","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"zenmux","id":"deepseek/deepseek-v3.2-exp","n":"DeepSeek-V3.2-Exp","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.22,"co":0.33,"ctx":163000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-09-29","lu":"2025-09-29"},{"p":"zenmux","id":"deepseek/deepseek-chat","n":"DeepSeek-V3.2 (Non-thinking Mode)","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":0.42,"cr":0.03,"ctx":128000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2025-01-01","rd":"2025-12-01","lu":"2025-12-01"},{"p":"zenmux","id":"deepseek/deepseek-v3.2","n":"DeepSeek V3.2","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":0.43,"ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-05","lu":"2025-12-05"},{"p":"zenmux","id":"minimax/minimax-m2.7-highspeed","n":"MiniMax M2.7 highspeed","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.611,"co":2.4439,"ctx":204800,"out":131070,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-03-20","lu":"2026-03-20"},{"p":"zenmux","id":"minimax/minimax-m2.5","n":"MiniMax M2.5","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-02-13","lu":"2026-02-13"},{"p":"zenmux","id":"minimax/minimax-m2.5-lightning","n":"MiniMax M2.5 highspeed","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":4.8,"cr":0.06,"cw":0.75,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-02-13","lu":"2026-02-13"},{"p":"zenmux","id":"minimax/minimax-m2.1","n":"MiniMax M2.1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"cw":0.38,"ctx":204000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-12-22","lu":"2025-12-22"},{"p":"zenmux","id":"minimax/minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.6,"co":2.4,"ctx":512000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"zenmux","id":"minimax/minimax-m2","n":"MiniMax M2","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"cw":0.38,"ctx":204000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2025-10-27","lu":"2025-10-27"},{"p":"zenmux","id":"minimax/minimax-m2.7","n":"MiniMax M2.7","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3055,"co":1.2219,"ctx":204800,"out":131070,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-03-20","lu":"2026-03-20"},{"p":"kenari","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0,"co":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"kenari","id":"kimi-k2-6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"kenari","id":"glm-5-1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0,"co":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"kenari","id":"gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"kenari","id":"grok-4-3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":0,"co":0,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"kenari","id":"qwen3-7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0,"co":0,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"kenari","id":"kimi-k2-7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"kenari","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0,"co":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"kenari","id":"deepseek-v4-flash:free","n":"DeepSeek V4 Flash (Free)","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0,"co":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"kenari","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":0,"co":0,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"kenari","id":"minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0,"co":0,"ctx":512000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"kenari","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":0,"co":0,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"kenari","id":"gpt-5-4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"kenari","id":"mimo-v2-5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0,"co":0,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"kenari","id":"gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"kenari","id":"gpt-5-5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":0,"co":0,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"kenari","id":"glm-5-2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"kenari","id":"grok-build-0-1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":0,"co":0,"ctx":256000,"out":256000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"kenari","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":0,"co":0,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"kenari","id":"gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT reasoning model for self-hosted agents and controllable deployments","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"kenari","id":"mimo-v2-5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0,"co":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"kenari","id":"deepseek-v4-pro:free","n":"DeepSeek V4 Pro (Free)","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0,"co":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"kenari","id":"gpt-image-2","n":"GPT-Image-2","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":272000,"out":16384,"mi":"ti","mo":"i","rd":"2026-04-21","lu":"2026-04-21"},{"p":"openai","id":"o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"openai","id":"text-embedding-3-large","n":"text-embedding-3-large","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.13,"co":0,"ctx":8191,"out":3072,"mi":"t","mo":"t","kn":"2024-01","rd":"2024-01-25","lu":"2024-01-25"},{"p":"openai","id":"gpt-5.2-pro","n":"GPT-5.2 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openai","id":"gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openai","id":"gpt-3.5-turbo","n":"GPT-3.5-turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"cr":0,"ctx":16385,"out":4096,"mi":"t","mo":"t","kn":"2021-09-01","rd":"2023-03-01","lu":"2023-11-06"},{"p":"openai","id":"gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"openai","id":"gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"openai","id":"gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"openai","id":"o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"openai","id":"o3-pro","n":"o3-pro","f":"o-pro","d":"High-effort o3 tier for difficult technical reasoning and careful answers","ci":20,"co":80,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-06-10","lu":"2025-06-10"},{"p":"openai","id":"chatgpt-image-latest","n":"chatgpt-image-latest","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"ti","mo":"ti","rd":"2025-12-16","lu":"2025-12-16"},{"p":"openai","id":"gpt-4o-2024-05-13","n":"GPT-4o (2024-05-13)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":15,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-05-13"},{"p":"openai","id":"gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"openai","id":"gpt-5-chat-latest","n":"GPT-5 Chat (latest)","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openai","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Codex GPT for repository edits, code review, and practical software agents","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openai","id":"gpt-5.3-codex-spark","n":"GPT-5.3 Codex Spark","f":"gpt-codex-spark","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"openai","id":"gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openai","id":"gpt-5.3-chat-latest","n":"GPT-5.3 Chat (latest)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-03","lu":"2026-03-03"},{"p":"openai","id":"gpt-4o-2024-08-06","n":"GPT-4o (2024-08-06)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-08-06","lu":"2024-08-06"},{"p":"openai","id":"text-embedding-ada-002","n":"text-embedding-ada-002","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":8192,"out":1536,"mi":"t","mo":"t","kn":"2022-12","rd":"2022-12-15","lu":"2022-12-15"},{"p":"openai","id":"o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"openai","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openai","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"openai","id":"text-embedding-3-small","n":"text-embedding-3-small","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.02,"co":0,"ctx":8191,"out":1536,"mi":"t","mo":"t","kn":"2024-01","rd":"2024-01-25","lu":"2024-01-25"},{"p":"openai","id":"gpt-5.1-codex-mini","n":"GPT-5.1 Codex mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openai","id":"gpt-5.1-chat-latest","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openai","id":"gpt-5.2-chat-latest","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openai","id":"o4-mini-deep-research","n":"o4-mini-deep-research","f":"o-mini","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-06-26","lu":"2024-06-26"},{"p":"openai","id":"gpt-image-1.5","n":"gpt-image-1.5","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"ti","mo":"ti","rd":"2025-11-25","lu":"2025-11-25"},{"p":"openai","id":"gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"openai","id":"gpt-4o-2024-11-20","n":"GPT-4o (2024-11-20)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-11-20","lu":"2024-11-20"},{"p":"openai","id":"o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"openai","id":"o1-pro","n":"o1-pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":150,"co":600,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2025-03-19","lu":"2025-03-19"},{"p":"openai","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"openai","id":"gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"openai","id":"gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"openai","id":"o3-deep-research","n":"o3-deep-research","f":"o","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":10,"co":40,"cr":2.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-06-26","lu":"2024-06-26"},{"p":"openai","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openai","id":"gpt-image-1","n":"gpt-image-1","f":"gpt-image","d":"OpenAI image model for production generation, edits, and brand-safe visual workflows","ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-04-24","lu":"2025-04-24"},{"p":"openai","id":"gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"openai","id":"gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"openai","id":"gpt-image-1-mini","n":"gpt-image-1-mini","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"ti","mo":"ti","rd":"2025-09-26","lu":"2025-09-26"},{"p":"openai","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openai","id":"gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"openai","id":"gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"openai","id":"gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"openai","id":"gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"openai","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openai","id":"gpt-image-2","n":"gpt-image-2","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":30,"cr":1.25,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2026-04-21","lu":"2026-04-21"},{"p":"openai","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openai","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"berget","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.99,"co":0.99,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2025-04-27","lu":"2025-04-27"},{"p":"berget","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.83,"co":3.85,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"berget","id":"google/gemma-4-31B-it","n":"Gemma 4 31B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.275,"co":0.55,"ctx":128000,"out":8192,"mi":"aitv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-04-02","lu":"2026-04-02"},{"p":"berget","id":"openai/gpt-oss-120b","n":"GPT-OSS-120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.22,"co":0.83,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-05","lu":"2025-08-05"},{"p":"berget","id":"mistralai/Mistral-Medium-3.5-128B","n":"Mistral Medium 3.5 128B","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":1.65,"co":5.5,"ctx":262144,"out":131072,"mi":"it","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-04","rd":"2026-04-29","lu":"2026-04-29"},{"p":"berget","id":"mistralai/Mistral-Small-3.2-24B-Instruct-2506","n":"Mistral Small 3.2 24B Instruct 2506","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.33,"co":0.33,"ctx":32000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2025-10-01","lu":"2025-10-01"},{"p":"berget","id":"zai-org/GLM-4.7","n":"GLM 4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.77,"co":2.75,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-01-19","lu":"2026-01-19"},{"p":"berget","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.54,"co":4.84,"ctx":524288,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"snowflake-cortex","id":"openai-gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"snowflake-cortex","id":"snowflake-llama3.3-70b","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"snowflake-cortex","id":"openai-gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"snowflake-cortex","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ctx":200000,"out":16384,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"snowflake-cortex","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"snowflake-cortex","id":"deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"snowflake-cortex","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"snowflake-cortex","id":"openai-gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"snowflake-cortex","id":"openai-gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"snowflake-cortex","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"snowflake-cortex","id":"openai-gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"snowflake-cortex","id":"claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ctx":200000,"out":16384,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"snowflake-cortex","id":"mistral-large2","n":"Mistral Large (latest)","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"snowflake-cortex","id":"openai-gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"snowflake-cortex","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ctx":1000000,"out":16384,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"snowflake-cortex","id":"openai-gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"snowflake-cortex","id":"gemini-3.1-pro","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"snowflake-cortex","id":"openai-gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ctx":272000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"tencent-token-plan","id":"hy3","n":"Hy3","f":"Hy","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-07-06","lu":"2026-07-06"},{"p":"github-models","id":"ai21-labs/ai21-jamba-1.5-mini","n":"AI21 Jamba 1.5 Mini","f":"jamba","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":256000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-03","rd":"2024-08-29","lu":"2024-08-29"},{"p":"github-models","id":"ai21-labs/ai21-jamba-1.5-large","n":"AI21 Jamba 1.5 Large","f":"jamba","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0,"co":0,"ctx":256000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-03","rd":"2024-08-29","lu":"2024-08-29"},{"p":"github-models","id":"core42/jais-30b-chat","n":"JAIS 30b Chat","f":"jais","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":8192,"out":2048,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-03","rd":"2023-08-30","lu":"2023-08-30"},{"p":"github-models","id":"xai/grok-3-mini","n":"Grok 3 Mini","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2024-12-09","lu":"2024-12-09"},{"p":"github-models","id":"xai/grok-3","n":"Grok 3","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2024-12-09","lu":"2024-12-09"},{"p":"github-models","id":"microsoft/phi-3.5-moe-instruct","n":"Phi-3.5-MoE instruct (128k)","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"github-models","id":"microsoft/phi-3-small-128k-instruct","n":"Phi-3-small instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"github-models","id":"microsoft/phi-3.5-mini-instruct","n":"Phi-3.5-mini instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"github-models","id":"microsoft/phi-3-medium-128k-instruct","n":"Phi-3-medium instruct (128k)","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"github-models","id":"microsoft/phi-3-small-8k-instruct","n":"Phi-3-small instruct (8k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":8192,"out":2048,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"github-models","id":"microsoft/phi-4-reasoning","n":"Phi-4-Reasoning","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"github-models","id":"microsoft/mai-ds-r1","n":"MAI-DS-R1","f":"mai","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":65536,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2025-01-20","lu":"2025-01-20"},{"p":"github-models","id":"microsoft/phi-4-mini-instruct","n":"Phi-4-mini-instruct","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"github-models","id":"microsoft/phi-4","n":"Phi-4","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":16000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"github-models","id":"microsoft/phi-3.5-vision-instruct","n":"Phi-3.5-vision instruct (128k)","f":"phi","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"github-models","id":"microsoft/phi-3-medium-4k-instruct","n":"Phi-3-medium instruct (4k)","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":4096,"out":1024,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"github-models","id":"microsoft/phi-4-multimodal-instruct","n":"Phi-4-multimodal-instruct","f":"phi","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"tia","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"github-models","id":"microsoft/phi-4-mini-reasoning","n":"Phi-4-mini-reasoning","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"github-models","id":"microsoft/phi-3-mini-128k-instruct","n":"Phi-3-mini instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"github-models","id":"microsoft/phi-3-mini-4k-instruct","n":"Phi-3-mini instruct (4k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":4096,"out":1024,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"github-models","id":"openai/o3","n":"OpenAI o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0,"co":0,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"kn":"2024-04","rd":"2025-01-31","lu":"2025-01-31"},{"p":"github-models","id":"openai/o1-mini","n":"OpenAI o1-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0,"co":0,"ctx":128000,"out":65536,"mi":"t","mo":"t","rs":1,"kn":"2023-10","rd":"2024-09-12","lu":"2024-12-17"},{"p":"github-models","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"tia","mo":"t","tc":1,"kn":"2023-10","rd":"2024-05-13","lu":"2024-05-13"},{"p":"github-models","id":"openai/o4-mini","n":"OpenAI o4-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0,"co":0,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"kn":"2024-04","rd":"2025-01-31","lu":"2025-01-31"},{"p":"github-models","id":"openai/o1-preview","n":"OpenAI o1-preview","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"kn":"2023-10","rd":"2024-09-12","lu":"2024-09-12"},{"p":"github-models","id":"openai/o3-mini","n":"OpenAI o3-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0,"co":0,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"kn":"2024-04","rd":"2025-01-31","lu":"2025-01-31"},{"p":"github-models","id":"openai/gpt-4.1-nano","n":"GPT-4.1-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"github-models","id":"openai/o1","n":"OpenAI o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0,"co":0,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"kn":"2023-10","rd":"2024-09-12","lu":"2024-12-17"},{"p":"github-models","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"github-models","id":"openai/gpt-4.1-mini","n":"GPT-4.1-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"github-models","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"tia","mo":"t","tc":1,"kn":"2023-10","rd":"2024-07-18","lu":"2024-07-18"},{"p":"github-models","id":"mistral-ai/mistral-small-2503","n":"Mistral Small 3.1","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09","rd":"2025-03-01","lu":"2025-03-01"},{"p":"github-models","id":"mistral-ai/mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-03","rd":"2024-07-18","lu":"2024-07-18"},{"p":"github-models","id":"mistral-ai/mistral-medium-2505","n":"Mistral Medium 3 (25.05)","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09","rd":"2025-05-01","lu":"2025-05-01"},{"p":"github-models","id":"mistral-ai/mistral-large-2411","n":"Mistral Large 24.11","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-09","rd":"2024-11-01","lu":"2024-11-01"},{"p":"github-models","id":"mistral-ai/ministral-3b","n":"Ministral 3B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-03","rd":"2024-10-22","lu":"2024-10-22"},{"p":"github-models","id":"mistral-ai/codestral-2501","n":"Codestral 25.01","f":"codestral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0,"co":0,"ctx":32000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-03","rd":"2025-01-01","lu":"2025-01-01"},{"p":"github-models","id":"cohere/cohere-command-a","n":"Cohere Command A","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-03","rd":"2024-11-01","lu":"2024-11-01"},{"p":"github-models","id":"cohere/cohere-command-r-08-2024","n":"Cohere Command R 08-2024","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-03","rd":"2024-08-01","lu":"2024-08-01"},{"p":"github-models","id":"cohere/cohere-command-r-plus","n":"Cohere Command R+","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-03","rd":"2024-04-04","lu":"2024-08-01"},{"p":"github-models","id":"cohere/cohere-command-r-plus-08-2024","n":"Cohere Command R+ 08-2024","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-03","rd":"2024-08-01","lu":"2024-08-01"},{"p":"github-models","id":"cohere/cohere-command-r","n":"Cohere Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-03","rd":"2024-03-11","lu":"2024-08-01"},{"p":"github-models","id":"meta/meta-llama-3-8b-instruct","n":"Meta-Llama-3-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":8192,"out":2048,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-04-18","lu":"2024-04-18"},{"p":"github-models","id":"meta/llama-3.2-11b-vision-instruct","n":"Llama-3.2-11B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"tia","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"github-models","id":"meta/meta-llama-3.1-405b-instruct","n":"Meta-Llama-3.1-405B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"github-models","id":"meta/llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout 17B 16E Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-31","lu":"2025-01-31"},{"p":"github-models","id":"meta/llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"github-models","id":"meta/meta-llama-3.1-70b-instruct","n":"Meta-Llama-3.1-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"github-models","id":"meta/meta-llama-3.1-8b-instruct","n":"Meta-Llama-3.1-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"github-models","id":"meta/llama-3.2-90b-vision-instruct","n":"Llama-3.2-90B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"tia","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"github-models","id":"meta/llama-4-maverick-17b-128e-instruct-fp8","n":"Llama 4 Maverick 17B 128E Instruct FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-31","lu":"2025-01-31"},{"p":"github-models","id":"meta/meta-llama-3-70b-instruct","n":"Meta-Llama-3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":8192,"out":2048,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-12","rd":"2024-04-18","lu":"2024-04-18"},{"p":"github-models","id":"deepseek/deepseek-r1-0528","n":"DeepSeek-R1-0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0,"co":0,"ctx":65536,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-05-28","lu":"2025-05-28"},{"p":"github-models","id":"deepseek/deepseek-v3-0324","n":"DeepSeek-V3-0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-03-24","lu":"2025-03-24"},{"p":"github-models","id":"deepseek/deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0,"co":0,"ctx":65536,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06","rd":"2025-01-20","lu":"2025-01-20"},{"p":"neuralwatt","id":"kimi-k2.5-fast","n":"Kimi K2.5 Fast","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.52,"co":2.59,"cr":0.13,"ctx":262128,"out":262128,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"neuralwatt","id":"kimi-k2.6-flex","n":"Kimi K2.6 Flex","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.345,"co":1.61,"cr":0.08625,"ctx":262128,"out":262128,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"neuralwatt","id":"glm-5.2-short-fast-flex","n":"GLM 5.2 Short Fast Flex","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.725,"co":2.25,"cr":0.18125,"ctx":199984,"out":199984,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"glm-5.2-flex","n":"GLM 5.2 Flex","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.725,"co":2.25,"cr":0.18125,"ctx":1048560,"out":1048560,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"glm-5.2","n":"GLM 5.2","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.45,"co":4.5,"cr":0.3625,"ctx":1048560,"out":1048560,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"glm-5.2-short-fast","n":"GLM 5.2 Short Fast","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":1.45,"co":4.5,"cr":0.3625,"ctx":199984,"out":199984,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"qwen3.5-397b-fast","n":"Qwen3.5 397B Fast","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.69,"co":4.14,"cr":0.1725,"ctx":262128,"out":262128,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"neuralwatt","id":"kimi-k2.6-fast","n":"Kimi K2.6 Fast","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.69,"co":3.22,"cr":0.1725,"ctx":262128,"out":262128,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"neuralwatt","id":"qwen3.6-35b-fast","n":"Qwen3.6 35B Fast","f":"qwen3.6","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.29,"co":1.15,"cr":0.0725,"ctx":131056,"out":131056,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"neuralwatt","id":"glm-5.2-short-flex","n":"GLM 5.2 Short Flex","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.725,"co":2.25,"cr":0.18125,"ctx":199984,"out":199984,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"glm-5.2-fast","n":"GLM 5.2 Fast","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":1.45,"co":4.5,"cr":0.3625,"ctx":1048560,"out":1048560,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"glm-5.2-short","n":"GLM 5.2 Short","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.45,"co":4.5,"cr":0.3625,"ctx":199984,"out":199984,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"neuralwatt","id":"kimi-k2.7-code-flex","n":"Kimi K2.7 Code Flex","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.475,"co":2,"cr":0.11875,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"neuralwatt","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.69,"co":3.22,"cr":0.1725,"ctx":262128,"out":262128,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"neuralwatt","id":"moonshotai/Kimi-K2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.52,"co":2.59,"cr":0.13,"ctx":262128,"out":262128,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"neuralwatt","id":"moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.2375,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"neuralwatt","id":"Qwen/Qwen3.6-35B-A3B","n":"Qwen3.6 35B A3B","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":1.15,"cr":0.0725,"ctx":131056,"out":131056,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"neuralwatt","id":"Qwen/Qwen3.5-397B-A17B-FP8","n":"Qwen3.5 397B A17B FP8","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.69,"co":4.14,"cr":0.1725,"ctx":262128,"out":262128,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"siliconflow-cn","id":"baidu/ERNIE-4.5-300B-A47B","n":"baidu/ERNIE-4.5-300B-A47B","f":"ernie","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.28,"co":1.1,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-02","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"ByteDance-Seed/Seed-OSS-36B-Instruct","n":"ByteDance-Seed/Seed-OSS-36B-Instruct","f":"seed","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.21,"co":0.57,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-04","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"stepfun-ai/Step-3.5-Flash","n":"stepfun-ai/Step-3.5-Flash","f":"step","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"ctx":262000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"siliconflow-cn","id":"inclusionAI/Ling-flash-2.0","n":"inclusionAI/Ling-flash-2.0","f":"ling","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.14,"co":0.57,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-18","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Pro/moonshotai/Kimi-K2.6","n":"Pro/moonshotai/Kimi-K2.6","f":"kimi","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"siliconflow-cn","id":"Pro/moonshotai/Kimi-K2.5","n":"Pro/moonshotai/Kimi-K2.5","f":"kimi","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.45,"co":2.25,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"siliconflow-cn","id":"Pro/zai-org/GLM-5","n":"Pro/zai-org/GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"ctx":205000,"out":205000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"siliconflow-cn","id":"Pro/zai-org/GLM-5.1","n":"Pro/zai-org/GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cw":0,"ctx":205000,"out":205000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-08","lu":"2026-04-08"},{"p":"siliconflow-cn","id":"Pro/deepseek-ai/DeepSeek-R1","n":"Pro/deepseek-ai/DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.18,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-28","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Pro/deepseek-ai/DeepSeek-V3.1-Terminus","n":"Pro/deepseek-ai/DeepSeek-V3.1-Terminus","f":"deepseek","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.27,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Pro/deepseek-ai/DeepSeek-V3.2","n":"Pro/deepseek-ai/DeepSeek-V3.2","f":"deepseek","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.27,"co":0.42,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-03","lu":"2025-12-03"},{"p":"siliconflow-cn","id":"Pro/deepseek-ai/DeepSeek-V3","n":"Pro/deepseek-ai/DeepSeek-V3","f":"deepseek","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.25,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","tc":1,"rd":"2024-12-26","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Pro/MiniMaxAI/MiniMax-M2.5","n":"Pro/MiniMaxAI/MiniMax-M2.5","f":"minimax","d":"Frontier MiniMax model for engineering, office tasks, and agentic reasoning","ci":0.3,"co":1.22,"ctx":192000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.6-35B-A3B","n":"Qwen/Qwen3.6-35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.23,"co":1.86,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-17","lu":"2026-04-17"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.5-397B-A17B","n":"Qwen/Qwen3.5-397B-A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":1.74,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen/Qwen3-235B-A22B-Thinking-2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.13,"co":0.6,"ctx":262000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-07-28","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.5-122B-A10B","n":"Qwen/Qwen3.5-122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":2.32,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-26","lu":"2026-02-26"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.5-27B","n":"Qwen/Qwen3.5-27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.09,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-25","lu":"2026-02-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-8B","n":"Qwen/Qwen3-8B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.06,"co":0.06,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-30","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.5-4B","n":"Qwen/Qwen3.5-4B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-03-03","lu":"2026-03-03"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.5-9B","n":"Qwen/Qwen3.5-9B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.22,"co":1.74,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-03-03","lu":"2026-03-03"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-32B","n":"Qwen/Qwen3-32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.14,"co":0.57,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-30","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-14B","n":"Qwen/Qwen3-14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.07,"co":0.28,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-30","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3.5-35B-A3B","n":"Qwen/Qwen3.5-35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.23,"co":1.86,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-25","lu":"2026-02-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-32B-Thinking","n":"Qwen/Qwen3-VL-32B-Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-21","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-30B-A3B-Instruct","n":"Qwen/Qwen3-VL-30B-A3B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":1,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-05","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen2.5-72B-Instruct","n":"Qwen/Qwen2.5-72B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.59,"co":0.59,"ctx":33000,"out":4000,"mi":"t","mo":"t","tc":1,"rd":"2024-09-18","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-32B-Instruct","n":"Qwen/Qwen3-VL-32B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.6,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-21","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-235B-A22B-Thinking","n":"Qwen/Qwen3-VL-235B-A22B-Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.45,"co":3.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-04","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-8B-Instruct","n":"Qwen/Qwen3-VL-8B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.18,"co":0.68,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-15","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-30B-A3B-Thinking","n":"Qwen/Qwen3-VL-30B-A3B-Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":1,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-11","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-30B-A3B-Instruct-2507","n":"Qwen/Qwen3-30B-A3B-Instruct-2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.3,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-30","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","n":"Qwen/Qwen3-Coder-30B-A3B-Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.07,"co":0.28,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-01","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","n":"Qwen/Qwen3-Coder-480B-A35B-Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.25,"co":1,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-31","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen3-VL-235B-A22B-Instruct","n":"Qwen/Qwen3-VL-235B-A22B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-04","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"Qwen/Qwen2.5-7B-Instruct","n":"Qwen/Qwen2.5-7B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.05,"co":0.05,"ctx":33000,"out":4000,"mi":"t","mo":"t","tc":1,"rd":"2024-09-18","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"PaddlePaddle/PaddleOCR-VL-1.5","n":"PaddlePaddle/PaddleOCR-VL-1.5","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0,"co":0,"ctx":16384,"out":16384,"mi":"ti","mo":"t","ow":1,"rd":"2026-01-29","lu":"2026-01-29"},{"p":"siliconflow-cn","id":"tencent/Hunyuan-A13B-Instruct","n":"tencent/Hunyuan-A13B-Instruct","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.14,"co":0.57,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-06-30","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cw":0,"ctx":1049000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"siliconflow-cn","id":"zai-org/GLM-4.5-Air","n":"zai-org/GLM-4.5-Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.14,"co":0.86,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-28","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-R1","n":"deepseek-ai/DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.18,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-28","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-V3.1-Terminus","n":"deepseek-ai/DeepSeek-V3.1-Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2025-11-25"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-V4-Flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.003,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-V4-Pro","n":"deepseek-ai/DeepSeek-V4-Pro","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.74,"co":3.48,"cr":0.145,"ctx":1049000,"out":393000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-V3.2","n":"deepseek-ai/DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.42,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-03","lu":"2025-12-03"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-OCR","n":"deepseek-ai/DeepSeek-OCR","d":"OCR model for extracting structured text from documents and screenshots","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-10-20","lu":"2025-10-20"},{"p":"siliconflow-cn","id":"deepseek-ai/DeepSeek-V3","n":"deepseek-ai/DeepSeek-V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.25,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","tc":1,"rd":"2024-12-26","lu":"2025-11-25"},{"p":"merge-gateway","id":"xai/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"merge-gateway","id":"xai/grok-4.20-0309-reasoning","n":"Grok 4.20 (Reasoning)","f":"grok","d":"Reasoning Grok for document-heavy analysis and long-horizon tool use","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"merge-gateway","id":"moonshotai/kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"merge-gateway","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Thinking Kimi model for slower research passes, planning, and hard technical questions","ci":0.6,"co":2.5,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"merge-gateway","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.6,"co":3,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"merge-gateway","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"merge-gateway","id":"moonshotai/kimi-k2.7-code-highspeed","n":"Kimi K2.7 Code Highspeed","f":"kimi-k2","d":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","ci":1.9,"co":8,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"merge-gateway","id":"mistral/codestral-latest","n":"Codestral (latest)","f":"codestral","d":"Mistral code model for completions, refactors, and developer IDE workflows","ci":0.3,"co":0.9,"ctx":256000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-05-29","lu":"2025-01-04"},{"p":"merge-gateway","id":"mistral/mistral-large-latest","n":"Mistral Large (latest)","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0.5,"co":1.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"merge-gateway","id":"mistral/devstral-small-2507","n":"Devstral Small","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.1,"co":0.3,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-05","rd":"2025-07-10","lu":"2025-07-10"},{"p":"merge-gateway","id":"mistral/pixtral-large-latest","n":"Pixtral Large (latest)","f":"pixtral","d":"Mistral's larger vision model for document-heavy image understanding and chat","ci":2,"co":6,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2024-11-04"},{"p":"merge-gateway","id":"mistral/mistral-medium-latest","n":"Mistral Medium (latest)","f":"mistral-medium","d":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"merge-gateway","id":"mistral/mistral-small-latest","n":"Mistral Small (latest)","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.15,"co":0.6,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-03-16","lu":"2026-03-16"},{"p":"merge-gateway","id":"mistral/mistral-medium-2505","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":131072,"out":131072,"mi":"ti","mo":"t","tc":1,"kn":"2025-05","rd":"2025-05-07","lu":"2025-05-07"},{"p":"merge-gateway","id":"mistral/mistral-large-2411","n":"Mistral Large 2.1","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-18","lu":"2024-11-18"},{"p":"merge-gateway","id":"mistral/magistral-medium-latest","n":"Magistral Medium (latest)","f":"magistral-medium","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":2,"co":5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-03-17","lu":"2025-03-20"},{"p":"merge-gateway","id":"mistral/devstral-medium-latest","n":"Devstral 2 (latest)","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-02","lu":"2025-12-02"},{"p":"merge-gateway","id":"mistral/devstral-2512","n":"Devstral 2","f":"devstral","d":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"merge-gateway","id":"mistral/mistral-large-2512","n":"Mistral Large 3","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ci":0.5,"co":1.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"merge-gateway","id":"mistral/devstral-medium-2507","n":"Devstral Medium","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":2,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"kn":"2025-05","rd":"2025-07-10","lu":"2025-07-10"},{"p":"merge-gateway","id":"google/gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"merge-gateway","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"merge-gateway","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"merge-gateway","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"merge-gateway","id":"google/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"merge-gateway","id":"google/gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"merge-gateway","id":"google/gemini-flash-lite-latest","n":"Gemini Flash-Lite Latest","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"merge-gateway","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"merge-gateway","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"merge-gateway","id":"google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"merge-gateway","id":"google/gemini-3-pro-preview","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"merge-gateway","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"merge-gateway","id":"google/gemini-flash-latest","n":"Gemini Flash Latest","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"merge-gateway","id":"google/gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"merge-gateway","id":"openai/o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"merge-gateway","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"merge-gateway","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"merge-gateway","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"merge-gateway","id":"openai/gpt-4o-2024-05-13","n":"GPT-4o (2024-05-13)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":15,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-05-13"},{"p":"merge-gateway","id":"openai/gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"merge-gateway","id":"openai/gpt-5-chat-latest","n":"GPT-5 Chat (latest)","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"merge-gateway","id":"openai/gpt-5.3-chat-latest","n":"GPT-5.3 Chat (latest)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-03","lu":"2026-03-03"},{"p":"merge-gateway","id":"openai/gpt-4o-2024-08-06","n":"GPT-4o (2024-08-06)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-08-06","lu":"2024-08-06"},{"p":"merge-gateway","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"merge-gateway","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"merge-gateway","id":"openai/gpt-5.1-chat-latest","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"merge-gateway","id":"openai/gpt-5.2-chat-latest","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"merge-gateway","id":"openai/gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"merge-gateway","id":"openai/gpt-4o-2024-11-20","n":"GPT-4o (2024-11-20)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-11-20","lu":"2024-11-20"},{"p":"merge-gateway","id":"openai/o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"merge-gateway","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"merge-gateway","id":"openai/gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"merge-gateway","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"merge-gateway","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"merge-gateway","id":"openai/gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"merge-gateway","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"merge-gateway","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"merge-gateway","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"merge-gateway","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"merge-gateway","id":"zai/glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"merge-gateway","id":"zai/glm-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"merge-gateway","id":"zai/glm-4.7-flashx","n":"GLM-4.7-FlashX","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"cr":0.01,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"merge-gateway","id":"zai/glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"merge-gateway","id":"zai/glm-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"merge-gateway","id":"zai/glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"merge-gateway","id":"zai/glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.2,"co":1.1,"cr":0.03,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"merge-gateway","id":"zai/glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":1,"co":3.2,"cr":0.2,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"merge-gateway","id":"zai/glm-5-turbo","n":"GLM-5-Turbo","f":"glm","d":"Faster GLM-5 lane for coding agents that need lower latency","ci":1.2,"co":4,"cr":0.24,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"merge-gateway","id":"anthropic/claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"merge-gateway","id":"anthropic/claude-opus-4-1-20250805","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"merge-gateway","id":"anthropic/claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"merge-gateway","id":"anthropic/claude-opus-4-5-20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-01","lu":"2025-11-01"},{"p":"merge-gateway","id":"anthropic/claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"merge-gateway","id":"anthropic/claude-opus-4-20250514","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"merge-gateway","id":"anthropic/claude-sonnet-4-20250514","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"merge-gateway","id":"anthropic/claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"merge-gateway","id":"anthropic/claude-sonnet-4-5-20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"merge-gateway","id":"anthropic/claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"merge-gateway","id":"cohere/command-a-03-2025","n":"Command A","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-03-13","lu":"2025-03-13"},{"p":"merge-gateway","id":"cohere/command-r-08-2024","n":"Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.15,"co":0.6,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"merge-gateway","id":"cohere/command-r7b-12-2024","n":"Command R7B","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.0375,"co":0.15,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-12-02","lu":"2024-12-02"},{"p":"merge-gateway","id":"cohere/command-r-plus-08-2024","n":"Command R+","f":"command-r","d":"Cohere's RAG workhorse for long-context enterprise search and tool use","ci":2.5,"co":10,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"merge-gateway","id":"alibaba/qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":1.65,"co":4.95,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"merge-gateway","id":"alibaba/qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.5,"co":3,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"merge-gateway","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"merge-gateway","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"merge-gateway","id":"minimax/minimax-m2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"merge-gateway","id":"minimax/minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"merge-gateway","id":"minimax/minimax-m2.1","n":"MiniMax-M2.1","f":"minimax","d":"Earlier MiniMax agent model for practical coding and productivity tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"merge-gateway","id":"minimax/minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.6,"co":2.4,"ctx":512000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"merge-gateway","id":"minimax/minimax-m2","n":"MiniMax-M2","f":"minimax","d":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","ci":0.3,"co":1.2,"ctx":196608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"merge-gateway","id":"minimax/minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"merge-gateway","id":"minimax/minimax-m2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"qihang-ai","id":"claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.14,"co":0.71,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-10-01","lu":"2025-10-01"},{"p":"qihang-ai","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.09,"co":0.71,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"qihang-ai","id":"claude-opus-4-5-20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":0.71,"co":3.57,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2025-11-01","lu":"2025-11-01"},{"p":"qihang-ai","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":0.25,"co":2,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"qihang-ai","id":"claude-sonnet-4-5-20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":0.43,"co":2.14,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"qihang-ai","id":"gemini-3-pro-preview","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":0.57,"co":3.43,"ctx":1000000,"out":65000,"mi":"tiav","mo":"t","rs":1,"tc":1,"kn":"2025-11","rd":"2025-11-19","lu":"2025-11-19"},{"p":"qihang-ai","id":"gpt-5-mini","n":"GPT-5-Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.04,"co":0.29,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"qihang-ai","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.07,"co":0.43,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"qihang-ai","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.14,"co":1.14,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2.5-tts","n":"MiMo-V2.5-TTS","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2-pro","n":"MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2-tts","n":"MiMo-V2-TTS","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2-omni","n":"MiMo-V2-Omni","f":"mimo","d":"MiMo omni model for text, image, video, audio, and agents","ci":0,"co":0,"cr":0,"ctx":262144,"out":131072,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2.5-tts-voicedesign","n":"MiMo-V2.5-TTS-VoiceDesign","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-ams","id":"mimo-v2.5-tts-voiceclone","n":"MiMo-V2.5-TTS-VoiceClone","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"modelscope","id":"Qwen/Qwen3-30B-A3B-Thinking-2507","n":"Qwen3 30B A3B Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-30","lu":"2025-07-30"},{"p":"modelscope","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen3-235B-A22B-Thinking-2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0,"co":0,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-25","lu":"2025-07-25"},{"p":"modelscope","id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","n":"Qwen3 Coder 30B A3B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-31","lu":"2025-07-31"},{"p":"modelscope","id":"Qwen/Qwen3-30B-A3B-Instruct-2507","n":"Qwen3 30B A3B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-30","lu":"2025-07-30"},{"p":"modelscope","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":262144,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04-28","lu":"2025-07-21"},{"p":"modelscope","id":"ZhipuAI/GLM-4.6","n":"GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"ctx":202752,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-09-30","lu":"2025-09-30"},{"p":"modelscope","id":"ZhipuAI/GLM-4.5","n":"GLM-4.5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"groq","id":"llama-3.3-70b-versatile","n":"Llama 3.3 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.59,"co":0.79,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"groq","id":"llama-3.1-8b-instant","n":"Llama 3.1 8B","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":0.05,"co":0.08,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"groq","id":"whisper-large-v3-turbo","n":"Whisper Large V3 Turbo","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ctx":0,"out":0,"mi":"a","mo":"t","ow":1,"rd":"2024-10-01","lu":"2024-10-01"},{"p":"groq","id":"whisper-large-v3","n":"Whisper","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ctx":0,"out":0,"mi":"a","mo":"t","ow":1,"rd":"2023-09-01","lu":"2025-09-05"},{"p":"groq","id":"meta-llama/llama-prompt-guard-2-86m","n":"Prompt Guard 2 86M","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.04,"co":0.04,"ctx":512,"out":512,"mi":"t","mo":"t","ow":1,"rd":"2025-05-29","lu":"2025-05-29"},{"p":"groq","id":"meta-llama/llama-prompt-guard-2-22m","n":"Llama Prompt Guard 2 22M","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.03,"co":0.03,"ctx":512,"out":512,"mi":"t","mo":"t","ow":1,"rd":"2025-05-29","lu":"2025-05-29"},{"p":"groq","id":"meta-llama/llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout 17B 16E","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.11,"co":0.34,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"groq","id":"openai/gpt-oss-safeguard-20b","n":"Safety GPT OSS 20B","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.075,"co":0.3,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-29","lu":"2026-06-29"},{"p":"groq","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.15,"co":0.6,"cr":0.075,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-10-21"},{"p":"groq","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.075,"co":0.3,"cr":0.0375,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-09-25"},{"p":"groq","id":"canopylabs/orpheus-v1-english","n":"Canopy Labs Orpheus V1 English","f":"canopylabs","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":4000,"out":50000,"mi":"t","mo":"a","rd":"2025-12-19","lu":"2025-12-19"},{"p":"groq","id":"canopylabs/orpheus-arabic-saudi","n":"Canopy Labs Orpheus Arabic Saudi","f":"canopylabs","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":4000,"out":50000,"mi":"t","mo":"a","rd":"2025-12-16","lu":"2025-12-16"},{"p":"groq","id":"groq/compound","n":"Compound","f":"groq","d":"General-purpose chat model for instruction following, writing, and analysis","ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-09-04","lu":"2025-09-04"},{"p":"groq","id":"groq/compound-mini","n":"Compound Mini","f":"groq","d":"Efficient model for low-latency assistance, extraction, and routine automation","ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-09-04","lu":"2025-09-04"},{"p":"groq","id":"qwen/qwen3-32b","n":"Qwen3-32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.29,"co":0.59,"ctx":131072,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-11","lu":"2025-06-12"},{"p":"mixlayer","id":"qwen/qwen3.5-27b","n":"Qwen3.5 27B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.3,"co":2.4,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"mixlayer","id":"qwen/qwen3.5-35b-a3b","n":"Qwen3.5 35B A3B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.25,"co":1.3,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"mixlayer","id":"qwen/qwen3.5-9b","n":"Qwen3.5 9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.4,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"mixlayer","id":"qwen/qwen3.5-397b-a17b","n":"Qwen3.5 397B A17B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.6,"co":3.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"mixlayer","id":"qwen/qwen3.5-122b-a10b","n":"Qwen3.5 122B A10B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.4,"co":3.2,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"orcarouter","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":2.5,"co":15,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"orcarouter","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"orcarouter","id":"google/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.13,"co":0.38,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"orcarouter","id":"google/gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":4,"co":18,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"orcarouter","id":"google/gemini-flash-lite-latest","n":"Gemini Flash-Lite Latest","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"orcarouter","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"orcarouter","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":4,"co":18,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"orcarouter","id":"google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.06,"co":0.33,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"orcarouter","id":"google/gemini-3-pro-preview","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":4,"co":18,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"orcarouter","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"orcarouter","id":"google/gemini-flash-latest","n":"Gemini Flash Latest","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.075,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"orcarouter","id":"google/gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"orcarouter","id":"z-ai/glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"orcarouter","id":"z-ai/glm-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"orcarouter","id":"z-ai/glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"orcarouter","id":"z-ai/glm-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"orcarouter","id":"z-ai/glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.2,"co":1.1,"cr":0.03,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"orcarouter","id":"z-ai/glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":1,"co":3.2,"cr":0.2,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"orcarouter","id":"openai/gpt-5.2-pro","n":"GPT-5.2 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"orcarouter","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"orcarouter","id":"openai/gpt-3.5-turbo","n":"GPT-3.5-turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"cr":0,"ctx":16385,"out":4096,"mi":"t","mo":"t","kn":"2021-09-01","rd":"2023-03-01","lu":"2023-11-06"},{"p":"orcarouter","id":"openai/gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"orcarouter","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"orcarouter","id":"openai/gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"orcarouter","id":"openai/gpt-4o-2024-05-13","n":"GPT-4o (2024-05-13)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":15,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-05-13"},{"p":"orcarouter","id":"openai/gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"orcarouter","id":"openai/gpt-5-chat-latest","n":"GPT-5 Chat (latest)","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"orcarouter","id":"openai/gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Codex GPT for repository edits, code review, and practical software agents","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"orcarouter","id":"openai/gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"orcarouter","id":"openai/gpt-5.3-chat-latest","n":"GPT-5.3 Chat (latest)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-03","lu":"2026-03-03"},{"p":"orcarouter","id":"openai/gpt-4o-2024-08-06","n":"GPT-4o (2024-08-06)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-08-06","lu":"2024-08-06"},{"p":"orcarouter","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"orcarouter","id":"openai/gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"orcarouter","id":"openai/gpt-5.1-codex-mini","n":"GPT-5.1 Codex mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"orcarouter","id":"openai/gpt-5.1-chat-latest","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Chat-tuned GPT-5.1 for polished assistants, writing, and product conversations","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"orcarouter","id":"openai/gpt-5.2-chat-latest","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"orcarouter","id":"openai/gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"orcarouter","id":"openai/gpt-4o-2024-11-20","n":"GPT-4o (2024-11-20)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-11-20","lu":"2024-11-20"},{"p":"orcarouter","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":5,"co":22.5,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"orcarouter","id":"openai/gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"orcarouter","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"orcarouter","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"orcarouter","id":"openai/gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"orcarouter","id":"openai/gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"orcarouter","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"orcarouter","id":"openai/gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":60,"co":270,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"orcarouter","id":"openai/gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"orcarouter","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"orcarouter","id":"openai/gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"orcarouter","id":"openai/gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"orcarouter","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"orcarouter","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"orcarouter","id":"kimi/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"orcarouter","id":"kimi/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"orcarouter","id":"orcarouter/auto","n":"OrcaRouter Auto","f":"auto","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-01-01","lu":"2026-05-14"},{"p":"orcarouter","id":"anthropic/claude-sonnet-4.5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"orcarouter","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"orcarouter","id":"anthropic/claude-haiku-4.5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"orcarouter","id":"anthropic/claude-opus-4.7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"orcarouter","id":"anthropic/claude-opus-4.1","n":"Claude Opus 4.1 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"orcarouter","id":"anthropic/claude-opus-4.5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"orcarouter","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"orcarouter","id":"anthropic/claude-opus-4","n":"Claude Opus 4 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"orcarouter","id":"anthropic/claude-opus-4.6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"orcarouter","id":"grok/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"orcarouter","id":"qwen/qwen3.6-35b-a3b","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.248,"co":1.485,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"orcarouter","id":"qwen/qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","ci":0.359,"co":1.434,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"orcarouter","id":"qwen/qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.115,"co":0.688,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"orcarouter","id":"qwen/qwen3.5-27b","n":"Qwen3.5 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.086,"co":0.688,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"orcarouter","id":"qwen/qwen3.5-35b-a3b","n":"Qwen3.5 35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.057,"co":0.459,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"orcarouter","id":"qwen/qwen3.5-397b-a17b","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.172,"co":1.032,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"orcarouter","id":"qwen/qwen3.5-122b-a10b","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.115,"co":0.917,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"orcarouter","id":"qwen/qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"orcarouter","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.19,"co":0.37,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"orcarouter","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.56,"co":1.12,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"orcarouter","id":"deepseek/deepseek-reasoner","n":"DeepSeek Reasoner","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.435,"co":0.87,"cr":0.028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"orcarouter","id":"deepseek/deepseek-chat","n":"DeepSeek Chat","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"orcarouter","id":"minimax/minimax-m2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"orcarouter","id":"minimax/minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"orcarouter","id":"minimax/minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"orcarouter","id":"minimax/minimax-m2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"helicone","id":"chatgpt-4o-latest","n":"OpenAI ChatGPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":20,"cr":2.5,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-08","rd":"2024-08-14","lu":"2024-08-14"},{"p":"helicone","id":"gpt-4.1-mini-2025-04-14","n":"OpenAI GPT-4.1 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.39999999999999997,"co":1.5999999999999999,"cr":0.09999999999999999,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2025-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"helicone","id":"deepseek-v3.1-terminus","n":"DeepSeek V3.1 Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"cr":0.21600000000000003,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-09","rd":"2025-09-22","lu":"2025-09-22"},{"p":"helicone","id":"claude-3.5-haiku","n":"Anthropic: Claude 3.5 Haiku","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.7999999999999999,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2024-10-22","lu":"2024-10-22"},{"p":"helicone","id":"llama-3.1-8b-instruct","n":"Meta Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.02,"co":0.049999999999999996,"ctx":16384,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-07-23","lu":"2024-07-23"},{"p":"helicone","id":"o3","n":"OpenAI o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"llama-prompt-guard-2-86m","n":"Meta Llama Prompt Guard 2 86M","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.01,"co":0.01,"ctx":512,"out":2,"mi":"t","mo":"t","kn":"2024-10","rd":"2024-10-01","lu":"2024-10-01"},{"p":"helicone","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3 Coder 30B A3B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.09999999999999999,"co":0.3,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"kn":"2025-07","rd":"2025-07-31","lu":"2025-07-31"},{"p":"helicone","id":"hermes-2-pro-llama-3-8b","n":"Hermes 2 Pro Llama 3 8B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.14,"co":0.14,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"kn":"2024-05","rd":"2024-05-27","lu":"2024-05-27"},{"p":"helicone","id":"deepseek-v3","n":"DeepSeek V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.56,"co":1.68,"cr":0.07,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2024-12-26","lu":"2024-12-26"},{"p":"helicone","id":"grok-code-fast-1","n":"xAI Grok Code Fast 1","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.19999999999999998,"co":1.5,"cr":0.02,"ctx":256000,"out":10000,"mi":"t","mo":"t","tc":1,"kn":"2024-08","rd":"2024-08-25","lu":"2024-08-25"},{"p":"helicone","id":"o1-mini","n":"OpenAI: o1-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":128000,"out":65536,"mi":"t","mo":"t","kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"deepseek-r1-distill-llama-70b","n":"DeepSeek R1 Distill Llama 70B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.03,"co":0.13,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-01-20","lu":"2025-01-20"},{"p":"helicone","id":"qwen3-32b","n":"Qwen3 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.29,"co":0.59,"ctx":131072,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-04-28","lu":"2025-04-28"},{"p":"helicone","id":"claude-sonnet-4","n":"Anthropic: Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.30000000000000004,"cw":3.75,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-05","rd":"2025-05-14","lu":"2025-05-14"},{"p":"helicone","id":"qwen3-next-80b-a3b-instruct","n":"Qwen3 Next 80B A3B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.14,"co":1.4,"ctx":262000,"out":16384,"mi":"tiv","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"llama-4-maverick","n":"Meta Llama 4 Maverick 17B 128E","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.15,"co":0.6,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":20,"co":40,"ctx":128000,"out":16400,"mi":"ti","mo":"t","kn":"2024-07","rd":"2024-07-18","lu":"2024-07-18"},{"p":"helicone","id":"llama-3.3-70b-versatile","n":"Meta Llama 3.3 70B Versatile","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.59,"co":0.7899999999999999,"ctx":131072,"out":32678,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"helicone","id":"gemma2-9b-it","n":"Google Gemma 2","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.01,"co":0.03,"ctx":8192,"out":8192,"mi":"t","mo":"t","kn":"2024-06","rd":"2024-06-25","lu":"2024-06-25"},{"p":"helicone","id":"gemini-2.5-pro","n":"Google Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.3125,"cw":1.25,"ctx":1048576,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-06-17","lu":"2025-06-17"},{"p":"helicone","id":"gpt-5","n":"OpenAI GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.12500000000000003,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"claude-haiku-4-5-20251001","n":"Anthropic: Claude 4.5 Haiku (20251001)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.09999999999999999,"cw":1.25,"ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2025-10","rd":"2025-10-01","lu":"2025-10-01"},{"p":"helicone","id":"claude-4.5-haiku","n":"Anthropic: Claude 4.5 Haiku","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.09999999999999999,"cw":1.25,"ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2025-10","rd":"2025-10-01","lu":"2025-10-01"},{"p":"helicone","id":"gpt-5-pro","n":"OpenAI: GPT-5 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":15,"co":120,"ctx":128000,"out":32768,"mi":"t","mo":"t","kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"gemini-2.5-flash","n":"Google Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"cw":0.3,"ctx":1048576,"out":65535,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-06","rd":"2025-06-17","lu":"2025-06-17"},{"p":"helicone","id":"gpt-4o","n":"OpenAI GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-05","rd":"2024-05-13","lu":"2024-05-13"},{"p":"helicone","id":"o4-mini","n":"OpenAI o4 Mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"llama-3.1-8b-instant","n":"Meta Llama 3.1 8B Instant","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":0.049999999999999996,"co":0.08,"ctx":131072,"out":32678,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-07-01","lu":"2024-07-01"},{"p":"helicone","id":"llama-prompt-guard-2-22m","n":"Meta Llama Prompt Guard 2 22M","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.01,"co":0.01,"ctx":512,"out":2,"mi":"t","mo":"t","kn":"2024-10","rd":"2024-10-01","lu":"2024-10-01"},{"p":"helicone","id":"o3-pro","n":"OpenAI o3 Pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":20,"co":80,"ctx":200000,"out":100000,"mi":"ti","mo":"t","tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"grok-3-mini","n":"xAI Grok 3 Mini","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.3,"co":0.5,"cr":0.075,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"claude-opus-4-1-20250805","n":"Anthropic: Claude Opus 4.1 (20250805)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08","rd":"2025-08-05","lu":"2025-08-05"},{"p":"helicone","id":"deepseek-tng-r1t2-chimera","n":"DeepSeek TNG R1T2 Chimera","f":"deepseek-thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.3,"co":1.2,"ctx":130000,"out":163840,"mi":"t","mo":"t","tc":1,"kn":"2025-07","rd":"2025-07-02","lu":"2025-07-02"},{"p":"helicone","id":"llama-3.3-70b-instruct","n":"Meta Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.13,"co":0.39,"ctx":128000,"out":16400,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"helicone","id":"sonar-reasoning-pro","n":"Perplexity Sonar Reasoning Pro","f":"sonar-reasoning","d":"Web-grounded reasoning model for multi-step research and cited answers","ci":2,"co":8,"ctx":127000,"out":4096,"mi":"t","mo":"t","rs":1,"kn":"2025-01","rd":"2025-01-27","lu":"2025-01-27"},{"p":"helicone","id":"grok-3","n":"xAI Grok 3","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"glm-4.6","n":"Zai GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.44999999999999996,"co":1.5,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2024-07-18","lu":"2024-07-18"},{"p":"helicone","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.48,"co":2,"ctx":256000,"out":262144,"mi":"t","mo":"t","tc":1,"kn":"2025-11","rd":"2025-11-06","lu":"2025-11-06"},{"p":"helicone","id":"grok-4-1-fast-non-reasoning","n":"xAI Grok 4.1 Fast Non-Reasoning","f":"grok","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.19999999999999998,"co":0.5,"cr":0.049999999999999996,"ctx":2000000,"out":30000,"mi":"ti","mo":"ti","tc":1,"kn":"2025-11","rd":"2025-11-17","lu":"2025-11-17"},{"p":"helicone","id":"qwen3-coder","n":"Qwen3 Coder 480B A35B Instruct Turbo","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.22,"co":0.95,"ctx":262144,"out":16384,"mi":"tiav","mo":"t","tc":1,"kn":"2025-07","rd":"2025-07-23","lu":"2025-07-23"},{"p":"helicone","id":"gpt-5-chat-latest","n":"OpenAI GPT-5 Chat Latest","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.12500000000000003,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-09","rd":"2024-09-30","lu":"2024-09-30"},{"p":"helicone","id":"gpt-5.1-codex","n":"OpenAI: GPT-5.1 Codex","f":"gpt-codex","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":1.25,"co":10,"cr":0.12500000000000003,"ctx":400000,"out":128000,"mi":"ti","mo":"ti","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"o3-mini","n":"OpenAI o3 Mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","tc":1,"kn":"2023-10","rd":"2023-10-01","lu":"2023-10-01"},{"p":"helicone","id":"mistral-large-2411","n":"Mistral-Large","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-07-24","lu":"2024-07-24"},{"p":"helicone","id":"gemini-2.5-flash-lite","n":"Google Gemini 2.5 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.09999999999999999,"co":0.39999999999999997,"cr":0.024999999999999998,"cw":0.09999999999999999,"ctx":1048576,"out":65535,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-07","rd":"2025-07-22","lu":"2025-07-22"},{"p":"helicone","id":"llama-guard-4","n":"Meta Llama Guard 4 12B","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.21,"co":0.21,"ctx":131072,"out":1024,"mi":"ti","mo":"t","kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"claude-opus-4-1","n":"Anthropic: Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08","rd":"2025-08-05","lu":"2025-08-05"},{"p":"helicone","id":"claude-3.7-sonnet","n":"Anthropic: Claude 3.7 Sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.30000000000000004,"cw":3.75,"ctx":200000,"out":64000,"mi":"ti","mo":"t","tc":1,"kn":"2025-02","rd":"2025-02-19","lu":"2025-02-19"},{"p":"helicone","id":"gpt-5.1-codex-mini","n":"OpenAI: GPT-5.1 Codex Mini","f":"gpt-codex","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.25,"co":2,"cr":0.024999999999999998,"ctx":400000,"out":128000,"mi":"ti","mo":"ti","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"gpt-5.1-chat-latest","n":"OpenAI GPT-5.1 Chat","f":"gpt-codex","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":1.25,"co":10,"cr":0.12500000000000003,"ctx":128000,"out":16384,"mi":"ti","mo":"ti","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"claude-3-haiku-20240307","n":"Anthropic: Claude 3 Haiku","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.25,"co":1.25,"cr":0.03,"cw":0.3,"ctx":200000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2024-03","rd":"2024-03-07","lu":"2024-03-07"},{"p":"helicone","id":"grok-4-fast-reasoning","n":"xAI: Grok 4 Fast Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.19999999999999998,"co":0.5,"cr":0.049999999999999996,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-09","rd":"2025-09-01","lu":"2025-09-01"},{"p":"helicone","id":"gpt-4.1-nano","n":"OpenAI GPT-4.1 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.09999999999999999,"co":0.39999999999999997,"cr":0.024999999999999998,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2025-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"helicone","id":"gpt-oss-120b","n":"OpenAI GPT-OSS 120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.04,"co":0.16,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"sonar","n":"Perplexity Sonar","f":"sonar","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":1,"co":1,"ctx":127000,"out":4096,"mi":"t","mo":"t","kn":"2025-01","rd":"2025-01-27","lu":"2025-01-27"},{"p":"helicone","id":"qwen2.5-coder-7b-fast","n":"Qwen2.5 Coder 7B fast","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.03,"co":0.09,"ctx":32000,"out":8192,"mi":"t","mo":"t","kn":"2024-09","rd":"2024-09-15","lu":"2024-09-15"},{"p":"helicone","id":"o1","n":"OpenAI: o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"t","mo":"t","kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"ernie-4.5-21b-a3b-thinking","n":"Baidu Ernie 4.5 21B A3B Thinking","f":"ernie","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.07,"co":0.28,"ctx":128000,"out":8000,"mi":"t","mo":"t","rs":1,"kn":"2025-03","rd":"2025-03-16","lu":"2025-03-16"},{"p":"helicone","id":"llama-4-scout","n":"Meta Llama 4 Scout 17B 16E","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.08,"co":0.3,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"sonar-pro","n":"Perplexity Sonar Pro","f":"sonar-pro","d":"Advanced Sonar search model for deeper research and cited synthesis","ci":3,"co":15,"ctx":200000,"out":4096,"mi":"t","mo":"t","kn":"2025-01","rd":"2025-01-27","lu":"2025-01-27"},{"p":"helicone","id":"gpt-4.1","n":"OpenAI GPT-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2025-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"helicone","id":"claude-sonnet-4-5-20250929","n":"Anthropic: Claude Sonnet 4.5 (20250929)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.30000000000000004,"cw":3.75,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-09","rd":"2025-09-29","lu":"2025-09-29"},{"p":"helicone","id":"deepseek-reasoner","n":"DeepSeek Reasoner","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.56,"co":1.68,"cr":0.07,"ctx":128000,"out":64000,"mi":"t","mo":"t","kn":"2025-01","rd":"2025-01-20","lu":"2025-01-20"},{"p":"helicone","id":"grok-4-1-fast-reasoning","n":"xAI Grok 4.1 Fast Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.19999999999999998,"co":0.5,"cr":0.049999999999999996,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-11","rd":"2025-11-17","lu":"2025-11-17"},{"p":"helicone","id":"gemini-3-pro-preview","n":"Google Gemini 3 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.19999999999999998,"ctx":1048576,"out":65536,"mi":"tiav","mo":"t","rs":1,"tc":1,"kn":"2025-11","rd":"2025-11-18","lu":"2025-11-18"},{"p":"helicone","id":"gpt-5-mini","n":"OpenAI GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.024999999999999998,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"gpt-4.1-mini","n":"OpenAI GPT-4.1 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.39999999999999997,"co":1.5999999999999999,"cr":0.09999999999999999,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2025-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"helicone","id":"sonar-reasoning","n":"Perplexity Sonar Reasoning","f":"sonar-reasoning","d":"Web-grounded reasoning model for multi-step research and cited answers","ci":1,"co":5,"ctx":127000,"out":4096,"mi":"t","mo":"t","rs":1,"kn":"2025-01","rd":"2025-01-27","lu":"2025-01-27"},{"p":"helicone","id":"sonar-deep-research","n":"Perplexity Sonar Deep Research","f":"sonar-deep-research","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":2,"co":8,"ctx":127000,"out":4096,"mi":"t","mo":"t","rs":1,"kn":"2025-01","rd":"2025-01-27","lu":"2025-01-27"},{"p":"helicone","id":"kimi-k2-0905","n":"Kimi K2 (09/05)","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.5,"co":2,"cr":0.39999999999999997,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2025-09","rd":"2025-09-05","lu":"2025-09-05"},{"p":"helicone","id":"gpt-5-nano","n":"OpenAI GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.049999999999999996,"co":0.39999999999999997,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"grok-4","n":"xAI Grok 4","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"ctx":256000,"out":256000,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-07-09","lu":"2024-07-09"},{"p":"helicone","id":"qwen3-235b-a22b-thinking","n":"Qwen3 235B A22B Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":2.9000000000000004,"ctx":262144,"out":81920,"mi":"tiv","mo":"t","rs":1,"kn":"2025-07","rd":"2025-07-25","lu":"2025-07-25"},{"p":"helicone","id":"claude-opus-4","n":"Anthropic: Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-05","rd":"2025-05-14","lu":"2025-05-14"},{"p":"helicone","id":"grok-4-fast-non-reasoning","n":"xAI Grok 4 Fast Non-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.19999999999999998,"co":0.5,"cr":0.049999999999999996,"ctx":2000000,"out":2000000,"mi":"tia","mo":"t","tc":1,"kn":"2025-09","rd":"2025-09-19","lu":"2025-09-19"},{"p":"helicone","id":"claude-4.5-opus","n":"Anthropic: Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-11","rd":"2025-11-24","lu":"2025-11-24"},{"p":"helicone","id":"qwen3-vl-235b-a22b-instruct","n":"Qwen3 VL 235B A22B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.5,"ctx":256000,"out":16384,"mi":"tiv","mo":"t","tc":1,"kn":"2025-09","rd":"2025-09-23","lu":"2025-09-23"},{"p":"helicone","id":"kimi-k2-0711","n":"Kimi K2 (07/11)","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.5700000000000001,"co":2.3,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"gemma-3-12b-it","n":"Google Gemma 3 12B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.049999999999999996,"co":0.09999999999999999,"ctx":131072,"out":8192,"mi":"ti","mo":"t","kn":"2024-12","rd":"2024-12-01","lu":"2024-12-01"},{"p":"helicone","id":"gpt-4o-mini","n":"OpenAI GPT-4o-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-07","rd":"2024-07-18","lu":"2024-07-18"},{"p":"helicone","id":"gpt-oss-20b","n":"OpenAI GPT-OSS 20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.049999999999999996,"co":0.19999999999999998,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2024-06-01","lu":"2024-06-01"},{"p":"helicone","id":"claude-3.5-sonnet-v2","n":"Anthropic: Claude 3.5 Sonnet v2","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.30000000000000004,"cw":3.75,"ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2024-10-22","lu":"2024-10-22"},{"p":"helicone","id":"qwen3-30b-a3b","n":"Qwen3 30B A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.08,"co":0.29,"ctx":41000,"out":41000,"mi":"ti","mo":"t","tc":1,"kn":"2025-06","rd":"2025-06-01","lu":"2025-06-01"},{"p":"helicone","id":"gpt-5-codex","n":"OpenAI: GPT-5 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.12500000000000003,"ctx":400000,"out":128000,"mi":"t","mo":"t","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"helicone","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.41,"ctx":163840,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-09","rd":"2025-09-22","lu":"2025-09-22"},{"p":"helicone","id":"claude-4.5-sonnet","n":"Anthropic: Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.30000000000000004,"cw":3.75,"ctx":200000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-09","rd":"2025-09-29","lu":"2025-09-29"},{"p":"helicone","id":"mistral-small","n":"Mistral Small 3.2","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.075,"co":0.2,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-06-20","lu":"2025-06-20"},{"p":"helicone","id":"llama-3.1-8b-instruct-turbo","n":"Meta Llama 3.1 8B Instruct Turbo","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":0.02,"co":0.03,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"kn":"2024-07","rd":"2024-07-23","lu":"2024-07-23"},{"p":"helicone","id":"gpt-5.1","n":"OpenAI GPT-5.1","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":1.25,"co":10,"cr":0.12500000000000003,"ctx":400000,"out":128000,"mi":"ti","mo":"ti","tc":1,"kn":"2025-01","rd":"2025-01-01","lu":"2025-01-01"},{"p":"zai","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"zai","id":"glm-4.5v","n":"GLM-4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"ctx":64000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-11","lu":"2025-08-11"},{"p":"zai","id":"glm-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zai","id":"glm-4.7-flashx","n":"GLM-4.7-FlashX","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"cr":0.01,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"zai","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"zai","id":"glm-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"zai","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"zai","id":"glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"zai","id":"glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":1.2,"co":4,"cr":0.24,"cw":0,"ctx":200000,"out":131072,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"zai","id":"glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.2,"co":1.1,"cr":0.03,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zai","id":"glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"zai","id":"glm-4.5-flash","n":"GLM-4.5-Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zai","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":1,"co":3.2,"cr":0.2,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"zai","id":"glm-5-turbo","n":"GLM-5-Turbo","f":"glm","d":"Faster GLM-5 lane for coding agents that need lower latency","ci":1.2,"co":4,"cr":0.24,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"nearai","id":"google/gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"nearai","id":"google/gemma-4-31B-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.13,"co":0.4,"cr":0.026,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"nearai","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"nearai","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"nearai","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"nearai","id":"google/gemini-3-pro","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":1.25,"co":15,"cr":0,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"nearai","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"nearai","id":"Qwen/Qwen3-Embedding-0.6B","n":"Qwen3 Embedding 0.6B","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.01,"co":0,"ctx":40960,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2025-06-03","lu":"2025-06-03"},{"p":"nearai","id":"Qwen/Qwen3-Reranker-0.6B","n":"Qwen3 Reranker 0.6B","f":"qwen","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0.01,"co":0.01,"ctx":40960,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2025-06-03","lu":"2025-06-03"},{"p":"nearai","id":"Qwen/Qwen3.6-35B-A3B-FP8","n":"Qwen 3.6 35B A3B FP8","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.17,"co":1.1,"cr":0.056,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"nearai","id":"Qwen/Qwen3.5-122B-A10B","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":3.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"nearai","id":"Qwen/Qwen3-30B-A3B-Instruct-2507","n":"Qwen3 30B-A3B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":0.55,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-29","lu":"2025-07-29"},{"p":"nearai","id":"Qwen/Qwen3-VL-30B-A3B-Instruct","n":"Qwen3-VL 30B-A3B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.15,"co":0.55,"ctx":256000,"out":32768,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-09-23","lu":"2025-09-23"},{"p":"nearai","id":"openai/o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"nearai","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"nearai","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"nearai","id":"openai/gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"nearai","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"nearai","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.8,"co":15.5,"cr":0.18,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"nearai","id":"openai/gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nearai","id":"openai/gpt-oss-120b","n":"GPT-OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.15,"co":0.55,"ctx":131000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"nearai","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"nearai","id":"openai/gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"nearai","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nearai","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"nearai","id":"openai/gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nearai","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"nearai","id":"openai/whisper-large-v3","n":"Whisper Large v3","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0.01,"co":0,"ctx":448,"out":448,"mi":"a","mo":"t","ow":1,"rd":"2023-11-06","lu":"2023-11-06"},{"p":"nearai","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"nearai","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"nearai","id":"anthropic/claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15.5,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"nearai","id":"anthropic/claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"nearai","id":"anthropic/claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"nearai","id":"anthropic/claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"nearai","id":"anthropic/claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"nearai","id":"zai-org/GLM-5.1-FP8","n":"GLM-5.1 FP8","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.85,"co":3.3,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"nearai","id":"black-forest-labs/FLUX.2-klein-4B","n":"FLUX.2 Klein 4B","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":1,"co":1,"ctx":128000,"out":128000,"mi":"ti","mo":"i","ow":1,"rd":"2026-01-14","lu":"2026-01-14"},{"p":"llmgateway","id":"claude-3-7-sonnet","n":"Claude 3.7 Sonnet","f":"claude","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-02-24","lu":"2025-02-24"},{"p":"llmgateway","id":"qwen-coder-plus","n":"Qwen Coder Plus","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.502,"co":1.004,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2024-09-18","lu":"2024-09-18"},{"p":"llmgateway","id":"mistral-large-latest","n":"Mistral Large (latest)","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":4,"co":12,"ctx":128000,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"llmgateway","id":"qwen3-vl-235b-a22b-thinking","n":"Qwen3 VL 235B A22B Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.5,"co":2,"ctx":131072,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-15","lu":"2025-09-15"},{"p":"llmgateway","id":"devstral-small-2507","n":"Devstral Small","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.1,"co":0.3,"ctx":131072,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-05","rd":"2025-07-10","lu":"2025-07-10"},{"p":"llmgateway","id":"qwen3-vl-30b-a3b-thinking","n":"Qwen3 VL 30B A3B Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1,"ctx":131072,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-02","lu":"2025-10-02"},{"p":"llmgateway","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1050000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"llmgateway","id":"qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Hosted Qwen coder for software agents, repo edits, and long-context code","ci":6,"co":60,"cr":1.2,"cw":7.5,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"llmgateway","id":"minimax-m2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"llmgateway","id":"qwen-plus","n":"Qwen Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.4,"co":1.2,"cr":0.08,"cw":0.5,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-09-11"},{"p":"llmgateway","id":"o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"llmgateway","id":"nemotron-3-ultra-550b","n":"Nemotron 3 Ultra 550B A55B","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.5,"co":2.5,"cr":0.15,"ctx":262144,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"llmgateway","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":228700,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"llmgateway","id":"grok-4-20-beta-0309-non-reasoning","n":"Grok 4.20 (Non-Reasoning)","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"cr":0.2,"ctx":2000000,"out":30000,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"llmgateway","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.38,"co":1.98,"cr":0.19,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"llmgateway","id":"gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":0.08333,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"llmgateway","id":"qwen3-235b-a22b-instruct-2507","n":"Qwen3 235B A22B Instruct (2507)","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.58,"ctx":262144,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-08","lu":"2025-07-08"},{"p":"llmgateway","id":"llama-3-70b-instruct","n":"Llama 3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.51,"co":0.74,"ctx":8192,"out":8000,"mi":"t","mo":"t","ow":1,"rd":"2024-04-18","lu":"2024-04-18"},{"p":"llmgateway","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3-Coder 30B-A3B Instruct","f":"qwen","d":"Smaller Qwen coder for efficient local agents and repo-level fixes","ci":0.07,"co":0.27,"ctx":262000,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"llmgateway","id":"seed-1-8-251228","n":"Seed 1.8 (251228)","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"cr":0.05,"ctx":256000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-18","lu":"2025-12-18"},{"p":"llmgateway","id":"kimi-k2","n":"Kimi K2","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.574,"co":2.294,"cr":0.5,"ctx":256000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-11","lu":"2025-07-11"},{"p":"llmgateway","id":"llama-3.1-70b-instruct","n":"Llama 3.1 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.72,"co":0.72,"ctx":128000,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2024-07-23","lu":"2024-07-23"},{"p":"llmgateway","id":"gpt-5.2-pro","n":"GPT-5.2 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"llmgateway","id":"minimax-m2.1","n":"MiniMax-M2.1","f":"minimax","d":"Earlier MiniMax agent model for practical coding and productivity tasks","ci":0.27,"co":1.1,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"llmgateway","id":"qwen3-32b","n":"Qwen3 32B","f":"qwen","d":"Dense open Qwen model for self-hosted chat, reasoning, and coding","ci":0.1,"co":0.3,"ctx":32768,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"llmgateway","id":"pixtral-large-latest","n":"Pixtral Large (latest)","f":"pixtral","d":"Mistral's larger vision model for document-heavy image understanding and chat","ci":4,"co":12,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2024-11-04"},{"p":"llmgateway","id":"glm-4-32b-0414-128k","n":"GLM-4 32B (0414-128k)","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.1,"co":0.1,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-04-14","lu":"2025-04-14"},{"p":"llmgateway","id":"seed-1-6-flash-250715","n":"Seed 1.6 Flash (250715)","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.07,"co":0.3,"cr":0.015,"ctx":256000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-26","lu":"2025-07-26"},{"p":"llmgateway","id":"qwen3-next-80b-a3b-instruct","n":"Qwen3-Next 80B-A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":1.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"llmgateway","id":"qwen3-vl-8b-instruct","n":"Qwen3 VL 8B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.08,"co":0.5,"ctx":131072,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-08-19","lu":"2025-08-19"},{"p":"llmgateway","id":"gpt-4o-mini-search-preview","n":"GPT-4o Mini Search Preview","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-10-01","lu":"2024-10-01"},{"p":"llmgateway","id":"glm-4.5v","n":"GLM-4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"cr":0.11,"ctx":128000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-11","lu":"2025-08-11"},{"p":"llmgateway","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.4,"co":1.6,"cr":0.08,"cw":0.5,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"llmgateway","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"llmgateway","id":"gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"llmgateway","id":"claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"llmgateway","id":"qwen3.6-35b-a3b","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.248,"co":1.485,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"llmgateway","id":"gpt-3.5-turbo","n":"GPT-3.5-turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"cr":0,"ctx":16385,"out":4096,"mi":"t","mo":"t","kn":"2021-09-01","rd":"2023-03-01","lu":"2023-11-06"},{"p":"llmgateway","id":"seed-1-6-250615","n":"Seed 1.6 (250615)","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"cr":0.05,"ctx":256000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-25","lu":"2025-06-25"},{"p":"llmgateway","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":1.25,"co":3.75,"cr":0.125,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"llmgateway","id":"glm-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"cw":0,"ctx":131000,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"llmgateway","id":"gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"llmgateway","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"llmgateway","id":"gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"llmgateway","id":"ministral-8b-2512","n":"Ministral 8B","f":"mistral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.15,"co":0.15,"ctx":262144,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"llmgateway","id":"gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"llmgateway","id":"o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"llmgateway","id":"qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","ci":0.845,"co":3.38,"cr":0.6,"cw":3.75,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"llmgateway","id":"mistral-small-2506","n":"Mistral Small 3.2","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-06-20","lu":"2025-06-20"},{"p":"llmgateway","id":"gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"cw":0.08333,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"llmgateway","id":"gemini-2.5-flash-lite-preview-09-2025","n":"Gemini 2.5 Flash Lite Preview (09-2025)","f":"gemini","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":1048576,"mi":"ti","mo":"t","tc":1,"rd":"2025-09-25","lu":"2025-09-25"},{"p":"llmgateway","id":"qwen-plus-latest","n":"Qwen Plus Latest","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":1.2,"cr":0.08,"cw":0.5,"ctx":1000000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2025-01-25","lu":"2025-01-25"},{"p":"llmgateway","id":"claude-opus-4-1-20250805","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"llmgateway","id":"gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.13,"co":0.38,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"llmgateway","id":"grok-4-3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.3125,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"llmgateway","id":"llama-4-scout-17b-instruct","n":"Llama 4 Scout 17B Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.18,"co":0.59,"ctx":131072,"out":2048,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-05","lu":"2025-04-05"},{"p":"llmgateway","id":"llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":0.13,"co":0.4,"ctx":131072,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"llmgateway","id":"sonar-reasoning-pro","n":"Sonar Reasoning Pro","f":"sonar-reasoning","d":"Web-grounded Sonar for multi-step research questions that need cited reasoning","ci":2,"co":8,"ctx":128000,"out":4096,"mi":"ti","mo":"t","rs":1,"kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"llmgateway","id":"grok-4-20-non-reasoning","n":"Grok 4.20 (Non-Reasoning)","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":30000,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"llmgateway","id":"glm-4.7-flashx","n":"GLM-4.7-FlashX","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"cr":0.01,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"llmgateway","id":"qwen3-30b-a3b-instruct-2507","n":"Qwen3 30B A3B Instruct (2507)","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.3,"ctx":262000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-08","lu":"2025-07-08"},{"p":"llmgateway","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"llmgateway","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.931,"co":2.93,"cr":0.173,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"llmgateway","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1050000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"llmgateway","id":"qwen3-next-80b-a3b-thinking","n":"Qwen3-Next 80B-A3B (Thinking)","f":"qwen","d":"Efficient Qwen thinking model for local reasoning, math, and coding agents","ci":0.15,"co":1.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"llmgateway","id":"glm-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.431,"co":2.007,"cr":0.11,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"llmgateway","id":"qwen35-397b-a17b","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.6,"co":3.6,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"llmgateway","id":"qwen3-235b-a22b-thinking-2507","n":"Qwen3 235B A22B Thinking (2507)","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.2,"co":0.6,"ctx":262000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-08","lu":"2025-07-08"},{"p":"llmgateway","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Thinking Kimi model for slower research passes, planning, and hard technical questions","ci":0.574,"co":2.294,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"llmgateway","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"llmgateway","id":"gemini-pro-latest","n":"Gemini Pro Latest","f":"gemini","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-27","lu":"2026-02-27"},{"p":"llmgateway","id":"glm-4.5-x","n":"GLM-4.5 X","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":2.2,"co":8.9,"cr":0.45,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"llmgateway","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"llmgateway","id":"claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"llmgateway","id":"grok-4-1-fast-non-reasoning","n":"Grok 4.1 Fast Non-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","tc":1,"rd":"2025-11-19","lu":"2025-11-19"},{"p":"llmgateway","id":"gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"llmgateway","id":"grok-4-20-beta-0309-reasoning","n":"Grok 4.20 (Reasoning)","f":"grok","d":"Reasoning Grok for document-heavy analysis and long-horizon tool use","ci":2,"co":6,"cr":0.2,"ctx":2000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"llmgateway","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.26,"co":3.96,"cr":0.234,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"llmgateway","id":"glm-4.5-airx","n":"GLM-4.5 AirX","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.1,"co":4.5,"cr":0.22,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"llmgateway","id":"gpt-5-chat-latest","n":"GPT-5 Chat (latest)","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"llmgateway","id":"ministral-3b-2512","n":"Ministral 3B","f":"mistral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"ctx":131072,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"llmgateway","id":"qwen2-5-vl-72b-instruct","n":"Qwen2.5-VL 72B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.13,"co":0.4,"ctx":32768,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"llmgateway","id":"glm-4.6v-flashx","n":"GLM-4.6V FlashX","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.04,"co":0.4,"cr":0.004,"ctx":128000,"out":16000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-08","lu":"2025-12-08"},{"p":"llmgateway","id":"minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.6,"co":2.4,"cr":0.12,"ctx":512000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"llmgateway","id":"qwen3-vl-plus","n":"Qwen3-VL Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1.6,"cr":0.04,"cw":0.25,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"llmgateway","id":"claude-opus-4-5-20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-01","lu":"2025-11-01"},{"p":"llmgateway","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Codex GPT for repository edits, code review, and practical software agents","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"llmgateway","id":"minimax-m2","n":"MiniMax-M2","f":"minimax","d":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","ci":0.2,"co":1,"cr":0.03,"ctx":196608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"llmgateway","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"llmgateway","id":"qwen3-max-2026-01-23","n":"Qwen3 Max (2026-01-23)","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.2,"co":6,"cr":0.24,"cw":1.5,"ctx":262144,"out":32800,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01-23","lu":"2026-01-23"},{"p":"llmgateway","id":"gpt-5.3-chat-latest","n":"GPT-5.3 Chat (latest)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-03-03","lu":"2026-03-03"},{"p":"llmgateway","id":"claude-3-opus","n":"Claude 3 Opus","f":"claude","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-03-04","lu":"2024-03-04"},{"p":"llmgateway","id":"qwen-omni-turbo","n":"Qwen-Omni Turbo","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.2,"co":0.8,"ctx":32768,"out":2048,"mi":"tiav","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-01-19","lu":"2025-03-26"},{"p":"llmgateway","id":"o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"llmgateway","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"llmgateway","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"llmgateway","id":"minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"llmgateway","id":"qwen-flash","n":"Qwen Flash","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.05,"co":0.4,"cr":0.01,"cw":0.0625,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"llmgateway","id":"qwen3-4b-fp8","n":"Qwen3 4B FP8","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.03,"co":0.03,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-28","lu":"2025-04-28"},{"p":"llmgateway","id":"gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"llmgateway","id":"qwen2-5-vl-32b-instruct","n":"Qwen2.5 VL 32B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.4,"co":4.2,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-03-15","lu":"2025-03-15"},{"p":"llmgateway","id":"gpt-5.1-codex-mini","n":"GPT-5.1 Codex mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"llmgateway","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.405,"co":1.98,"cr":0.225,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"llmgateway","id":"seed-1-6-250915","n":"Seed 1.6 (250915)","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"cr":0.05,"ctx":256000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-15","lu":"2025-09-15"},{"p":"llmgateway","id":"gpt-5.2-chat-latest","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"llmgateway","id":"minimax-text-01","n":"MiniMax Text 01","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.2,"co":1.1,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-01-15","lu":"2025-01-15"},{"p":"llmgateway","id":"gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1000000,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"llmgateway","id":"gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.25,"ctx":131072,"out":32766,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"llmgateway","id":"qwen-vl-max","n":"Qwen-VL Max","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.8,"co":3.2,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-04-08","lu":"2025-08-13"},{"p":"llmgateway","id":"llama-3-8b-instruct","n":"Llama 3 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.04,"co":0.04,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2025-04-03","lu":"2025-04-03"},{"p":"llmgateway","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"llmgateway","id":"sonar","n":"Sonar","f":"sonar","d":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","ci":1,"co":1,"ctx":130000,"out":4096,"mi":"t","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"llmgateway","id":"qwen-max","n":"Qwen Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.6,"co":6.4,"ctx":32768,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-04-03","lu":"2025-01-25"},{"p":"llmgateway","id":"claude-3-7-sonnet-20250219","n":"Claude Sonnet 3.7","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10-31","rd":"2025-02-19","lu":"2025-02-19"},{"p":"llmgateway","id":"o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"llmgateway","id":"minimax-m2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"llmgateway","id":"deepseek-v3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.56,"co":1.68,"cr":0.112,"ctx":128000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-21","lu":"2025-08-21"},{"p":"llmgateway","id":"ministral-14b-2512","n":"Ministral 14B","f":"mistral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"ctx":262144,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"llmgateway","id":"sonar-pro","n":"Sonar Pro","f":"sonar-pro","d":"Deeper Sonar search model with broader retrieval and stronger synthesis","ci":3,"co":15,"ctx":200000,"out":8192,"mi":"ti","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"llmgateway","id":"glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"cr":0.05,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"llmgateway","id":"claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"llmgateway","id":"claude-haiku-4-5-free","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":0,"co":0,"ctx":200000,"out":200000,"mi":"ti","mo":"t","tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"llmgateway","id":"minimax-m2.1-lightning","n":"MiniMax M2.1 Lightning","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.12,"co":0.48,"ctx":196608,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"llmgateway","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"llmgateway","id":"mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"llmgateway","id":"mimo-v2-omni","n":"MiMo-V2-Omni","f":"mimo","d":"MiMo omni model for text, image, video, audio, and agents","ci":0.4,"co":2,"cr":0.08,"ctx":256000,"out":131072,"mi":"tiavp","mo":"t","tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"llmgateway","id":"gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"llmgateway","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.4,"co":2.2,"cr":0.08,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"llmgateway","id":"gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1000000,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"llmgateway","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"llmgateway","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"llmgateway","id":"qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.108,"co":0.675,"cr":0.06,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"llmgateway","id":"llama-4-maverick-17b-instruct","n":"Llama 4 Maverick 17B Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.27,"co":0.85,"ctx":1048576,"out":2048,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-05","lu":"2025-04-05"},{"p":"llmgateway","id":"qwen3-coder-480b-a35b-instruct","n":"Qwen3-Coder 480B-A35B Instruct","f":"qwen","d":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","ci":0.3,"co":1.3,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"llmgateway","id":"devstral-2512","n":"Devstral 2","f":"devstral","d":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","ci":0.4,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"llmgateway","id":"claude-sonnet-4-5-20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"llmgateway","id":"gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.07,"co":0.34,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"llmgateway","id":"qwen3.5-9b","n":"Qwen3.5 9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.15,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"llmgateway","id":"qwq-plus","n":"QwQ Plus","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.8,"co":2.4,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-03-05","lu":"2025-03-05"},{"p":"llmgateway","id":"grok-4-1-fast-reasoning","n":"Grok 4.1 Fast Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-19","lu":"2025-11-19"},{"p":"llmgateway","id":"qwen3-vl-30b-a3b-instruct","n":"Qwen3 VL 30B A3B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.7,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-10-02","lu":"2025-10-02"},{"p":"llmgateway","id":"qwen3-vl-flash","n":"Qwen3 VL Flash","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.05,"co":0.4,"cr":0.01,"ctx":262144,"out":32000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-09","lu":"2025-10-09"},{"p":"llmgateway","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"llmgateway","id":"gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1000000,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"llmgateway","id":"llama-3.1-nemotron-ultra-253b","n":"Llama 3.1 Nemotron Ultra 253B","f":"nemotron","d":"Flagship Nemotron model for high-throughput reasoning and complex agents","ci":0.6,"co":1.8,"ctx":128000,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2025-04-07","lu":"2025-04-07"},{"p":"llmgateway","id":"gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"llmgateway","id":"qwen3-coder-flash","n":"Qwen3 Coder Flash","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.3,"co":1.5,"cr":0.06,"cw":0.375,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"llmgateway","id":"grok-build-0-1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"llmgateway","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"llmgateway","id":"grok-4","n":"Grok 4","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"ctx":256000,"out":256000,"mi":"ti","mo":"t","tc":1,"rd":"2025-07-09","lu":"2025-07-09"},{"p":"llmgateway","id":"llama-3.2-11b-instruct","n":"Llama 3.2 11B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.07,"co":0.33,"ctx":128000,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-09-25","lu":"2024-09-25"},{"p":"llmgateway","id":"gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"llmgateway","id":"glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.13,"co":0.85,"cr":0.025,"cw":0,"ctx":131000,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"llmgateway","id":"kimi-k2.7-code-highspeed","n":"Kimi K2.7 Code Highspeed","f":"kimi-k2","d":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","ci":1.9,"co":8,"cr":0.38,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"llmgateway","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"llmgateway","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"llmgateway","id":"glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0.06,"co":0.4,"cr":0.01,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"llmgateway","id":"qwen3-vl-235b-a22b-instruct","n":"Qwen3 VL 235B A22B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.5,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-09-15","lu":"2025-09-15"},{"p":"llmgateway","id":"mimo-v2-pro","n":"MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":1,"co":3,"cr":0.2,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"llmgateway","id":"gpt-4o-search-preview","n":"GPT-4o Search Preview","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-10-01","lu":"2024-10-01"},{"p":"llmgateway","id":"llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.03,"co":0.05,"ctx":32768,"out":32000,"mi":"t","mo":"t","ow":1,"rd":"2024-09-18","lu":"2024-09-18"},{"p":"llmgateway","id":"mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"llmgateway","id":"gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"llmgateway","id":"gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"llmgateway","id":"gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.04,"co":0.15,"ctx":131072,"out":32766,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"llmgateway","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.72,"co":2.3,"cr":0.144,"cw":0,"ctx":203000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"llmgateway","id":"qwen-max-latest","n":"Qwen Max Latest","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.6,"co":6.4,"ctx":32768,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2025-01-25","lu":"2025-01-25"},{"p":"llmgateway","id":"mistral-large-2512","n":"Mistral Large 3","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ci":0.5,"co":1.5,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"llmgateway","id":"qwen-turbo","n":"Qwen Turbo","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.05,"co":0.2,"ctx":1000000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2024-11-01","lu":"2025-04-28"},{"p":"llmgateway","id":"custom","n":"Custom Model","f":"auto","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"llmgateway","id":"grok-4-20-reasoning","n":"Grok 4.20 (Reasoning)","f":"grok","d":"Reasoning Grok for document-heavy analysis and long-horizon tool use","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-09","lu":"2026-03-09"},{"p":"llmgateway","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.26,"co":0.38,"cr":0.13,"ctx":163840,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"llmgateway","id":"qwen3.6-max-preview","n":"Qwen3.6 Max Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.3,"co":7.8,"cr":0.13,"cw":1.625,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-20","lu":"2026-04-20"},{"p":"llmgateway","id":"auto","n":"Auto Route","f":"auto","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"llmgateway","id":"codestral-2508","n":"Codestral","f":"mistral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.9,"ctx":256000,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2025-07-30","lu":"2025-07-30"},{"p":"llmgateway","id":"qwen3-235b-a22b-fp8","n":"Qwen3 235B A22B FP8","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.2,"co":0.8,"ctx":40960,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-28","lu":"2025-04-28"},{"p":"llmgateway","id":"fugu-ultra","n":"Fugu Ultra","d":"Quality-first multi-agent model for hard research, analysis, and competitions","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-22","lu":"2026-06-22"},{"p":"llmgateway","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"llmgateway","id":"qwen-vl-plus","n":"Qwen-VL Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.21,"co":0.64,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-08-15"},{"p":"llmgateway","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"llmgateway","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"llmgateway","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"alibaba-coding-plan-cn","id":"qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"alibaba-coding-plan-cn","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"alibaba-coding-plan-cn","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"alibaba-coding-plan-cn","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.5,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"alibaba-coding-plan-cn","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1875,"co":1.125,"cw":0.234375,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"alibaba-coding-plan-cn","id":"qwen3-max-2026-01-23","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2026-01-23","lu":"2026-01-23"},{"p":"alibaba-coding-plan-cn","id":"qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"alibaba-coding-plan-cn","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"alibaba-coding-plan-cn","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":196608,"out":24576,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-coding-plan-cn","id":"qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-03","lu":"2026-02-03"},{"p":"alibaba-coding-plan-cn","id":"glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"alibaba-coding-plan-cn","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"abacus","id":"o3","n":"o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2,"co":8,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"abacus","id":"route-llm","n":"Route LLM","f":"gpt","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":3,"co":15,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2024-01-01","lu":"2024-01-01"},{"p":"abacus","id":"grok-code-fast-1","n":"Grok Code Fast 1","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":1.5,"ctx":256000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-09-01","lu":"2025-09-01"},{"p":"abacus","id":"gpt-5.3-codex-xhigh","n":"GPT-5.3 Codex XHigh","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"abacus","id":"llama-3.3-70b-versatile","n":"Llama 3.3 70B Versatile","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.59,"co":0.79,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-12-06","lu":"2024-12-06"},{"p":"abacus","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-25","lu":"2025-03-25"},{"p":"abacus","id":"gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"abacus","id":"claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"abacus","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-05"},{"p":"abacus","id":"o4-mini","n":"o4-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"abacus","id":"qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.2,"co":6,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-28","lu":"2025-05-28"},{"p":"abacus","id":"o3-pro","n":"o3-pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":20,"co":40,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-06-10","lu":"2025-06-10"},{"p":"abacus","id":"claude-opus-4-1-20250805","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"abacus","id":"grok-4-1-fast-non-reasoning","n":"Grok 4.1 Fast (Non-Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"ctx":2000000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-11-17","lu":"2025-11-17"},{"p":"abacus","id":"claude-opus-4-5-20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-01","lu":"2025-11-01"},{"p":"abacus","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"abacus","id":"gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"abacus","id":"claude-opus-4-20250514","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-14","lu":"2025-05-14"},{"p":"abacus","id":"gpt-5.3-chat-latest","n":"GPT-5.3 Chat Latest","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"abacus","id":"o3-mini","n":"o3-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"abacus","id":"kimi-k2-turbo-preview","n":"Kimi K2 Turbo Preview","f":"kimi-k2","d":"Fast Kimi model for responsive chat, coding help, and agent loops","ci":0.15,"co":8,"ctx":256000,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-07-08","lu":"2025-07-08"},{"p":"abacus","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"abacus","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"abacus","id":"grok-4-0709","n":"Grok 4","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"ctx":256000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-07-09","lu":"2025-07-09"},{"p":"abacus","id":"claude-sonnet-4-20250514","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-14","lu":"2025-05-14"},{"p":"abacus","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"abacus","id":"gpt-5.1-chat-latest","n":"GPT-5.1 Chat Latest","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"abacus","id":"gpt-5.2-chat-latest","n":"GPT-5.2 Chat Latest","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2026-01-01","lu":"2026-01-01"},{"p":"abacus","id":"gpt-4.1-nano","n":"GPT-4.1 Nano","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"abacus","id":"gpt-4o-2024-11-20","n":"GPT-4o (2024-11-20)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"tia","mo":"t","tc":1,"kn":"2024-10","rd":"2024-11-20","lu":"2024-11-20"},{"p":"abacus","id":"claude-3-7-sonnet-20250219","n":"Claude Sonnet 3.7","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10-31","rd":"2025-02-19","lu":"2025-02-19"},{"p":"abacus","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"abacus","id":"gpt-4.1","n":"GPT-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"abacus","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"abacus","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"abacus","id":"qwen-2.5-coder-32b","n":"Qwen 2.5 Coder 32B","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.79,"co":0.79,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-11","lu":"2024-11-11"},{"p":"abacus","id":"claude-sonnet-4-5-20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"abacus","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"abacus","id":"gpt-4.1-mini","n":"GPT-4.1 Mini","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"abacus","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"abacus","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"abacus","id":"grok-4-fast-non-reasoning","n":"Grok 4 Fast (Non-Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"ctx":2000000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-07-09","lu":"2025-07-09"},{"p":"abacus","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"abacus","id":"gpt-4o-mini","n":"GPT-4o Mini","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-07-18","lu":"2024-07-18"},{"p":"abacus","id":"gpt-5-codex","n":"GPT-5 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"abacus","id":"gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":1,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"abacus","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"abacus","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"abacus","id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","n":"Llama 4 Maverick 17B 128E Instruct FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.14,"co":0.59,"ctx":1000000,"out":32768,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"abacus","id":"meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo","n":"Llama 3.1 405B Instruct Turbo","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":3.5,"co":3.5,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-23","lu":"2024-07-23"},{"p":"abacus","id":"meta-llama/Meta-Llama-3.1-8B-Instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.02,"co":0.05,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-23","lu":"2024-07-23"},{"p":"abacus","id":"Qwen/qwen3-coder-480b-a35b-instruct","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.29,"co":1.2,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-22","lu":"2025-07-22"},{"p":"abacus","id":"Qwen/QwQ-32B","n":"QwQ 32B","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.4,"co":0.4,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2024-11-28","lu":"2024-11-28"},{"p":"abacus","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.13,"co":0.6,"ctx":262144,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-01","lu":"2025-07-01"},{"p":"abacus","id":"Qwen/Qwen3-32B","n":"Qwen3 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.29,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"abacus","id":"Qwen/Qwen2.5-72B-Instruct","n":"Qwen 2.5 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.11,"co":0.38,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09-19","lu":"2024-09-19"},{"p":"abacus","id":"openai/gpt-oss-120b","n":"GPT-OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.08,"co":0.44,"ctx":128000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"abacus","id":"zai-org/glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-06-01","lu":"2025-06-01"},{"p":"abacus","id":"zai-org/glm-4.5","n":"GLM-4.5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"abacus","id":"zai-org/glm-4.6","n":"GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-03-01","lu":"2025-03-01"},{"p":"abacus","id":"zai-org/glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"abacus","id":"deepseek-ai/DeepSeek-R1","n":"DeepSeek R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":3,"co":7,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"abacus","id":"deepseek-ai/DeepSeek-V3.1-Terminus","n":"DeepSeek V3.1 Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-01","lu":"2025-06-01"},{"p":"abacus","id":"deepseek-ai/DeepSeek-V3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.4,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-15","lu":"2025-06-15"},{"p":"abacus","id":"deepseek/deepseek-v3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.55,"co":1.66,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"cloudferro-sherlock","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.92,"co":2.92,"ctx":70000,"out":70000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10-09","rd":"2024-12-06","lu":"2024-12-06"},{"p":"cloudferro-sherlock","id":"openai/gpt-oss-120b","n":"OpenAI GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":2.92,"co":2.92,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-28","lu":"2025-08-28"},{"p":"cloudferro-sherlock","id":"speakleash/Bielik-11B-v3.0-Instruct","n":"Bielik 11B v3.0 Instruct","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.67,"co":0.67,"ctx":32000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-03-13","lu":"2025-03-13"},{"p":"cloudferro-sherlock","id":"speakleash/Bielik-11B-v2.6-Instruct","n":"Bielik 11B v2.6 Instruct","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.67,"co":0.67,"ctx":32000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-03-13","lu":"2025-03-13"},{"p":"cloudferro-sherlock","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":196000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-03-05","lu":"2026-03-05"},{"p":"ollama-cloud","id":"deepseek-v4-flash","n":"deepseek-v4-flash","f":"deepseek-flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ctx":1048576,"out":1048576,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"ollama-cloud","id":"minimax-m2.5","n":"minimax-m2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"ollama-cloud","id":"devstral-small-2:24b","n":"devstral-small-2:24b","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-09","lu":"2026-01-19"},{"p":"ollama-cloud","id":"glm-4.7","n":"glm-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-22","lu":"2026-01-19"},{"p":"ollama-cloud","id":"cogito-2.1:671b","n":"cogito-2.1:671b","f":"cogito","d":"Legacy model retained for compatibility with older integrations","ctx":163840,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-19","lu":"2026-01-19"},{"p":"ollama-cloud","id":"minimax-m2.1","n":"minimax-m2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2026-01-19"},{"p":"ollama-cloud","id":"gpt-oss:120b","n":"gpt-oss:120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2026-01-19"},{"p":"ollama-cloud","id":"nemotron-3-nano:30b","n":"nemotron-3-nano:30b","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-15","lu":"2026-01-19"},{"p":"ollama-cloud","id":"ministral-3:8b","n":"ministral-3:8b","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ctx":262144,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2024-12-01","lu":"2026-01-19"},{"p":"ollama-cloud","id":"rnj-1:8b","n":"rnj-1:8b","f":"rnj","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ctx":32768,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-06","lu":"2026-01-19"},{"p":"ollama-cloud","id":"kimi-k2.7-code","n":"kimi-k2.7-code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"ollama-cloud","id":"glm-5.1","n":"glm-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-04-07"},{"p":"ollama-cloud","id":"deepseek-v4-pro","n":"deepseek-v4-pro","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ctx":1048576,"out":1048576,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"ollama-cloud","id":"glm-4.6","n":"glm-4.6","f":"glm","d":"Legacy model retained for compatibility with older integrations","ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-29","lu":"2026-01-19"},{"p":"ollama-cloud","id":"kimi-k2-thinking","n":"kimi-k2-thinking","f":"kimi-thinking","d":"Legacy model retained for compatibility with older integrations","ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2026-01-19"},{"p":"ollama-cloud","id":"nemotron-3-super","n":"nemotron-3-super","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-12"},{"p":"ollama-cloud","id":"ministral-3:14b","n":"ministral-3:14b","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ctx":262144,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2024-12-01","lu":"2026-01-19"},{"p":"ollama-cloud","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ctx":976000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"ollama-cloud","id":"minimax-m3","n":"minimax-m3","f":"minimax-m3","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ctx":512000,"out":131072,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-05-31","lu":"2026-05-31"},{"p":"ollama-cloud","id":"minimax-m2","n":"minimax-m2","f":"minimax","d":"Legacy model retained for compatibility with older integrations","ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-23","lu":"2026-01-19"},{"p":"ollama-cloud","id":"qwen3-next:80b","n":"qwen3-next:80b","f":"qwen","d":"Legacy model retained for compatibility with older integrations","ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-15","lu":"2026-01-19"},{"p":"ollama-cloud","id":"qwen3-coder:480b","n":"qwen3-coder:480b","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-22","lu":"2026-01-19"},{"p":"ollama-cloud","id":"kimi-k2:1t","n":"kimi-k2:1t","f":"kimi-k2","d":"Legacy model retained for compatibility with older integrations","ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-07-11","lu":"2026-01-19"},{"p":"ollama-cloud","id":"minimax-m2.7","n":"minimax-m2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"ollama-cloud","id":"deepseek-v3.1:671b","n":"deepseek-v3.1:671b","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-21","lu":"2026-01-19"},{"p":"ollama-cloud","id":"kimi-k2.5","n":"kimi-k2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"ollama-cloud","id":"gemma4:31b","n":"gemma4:31b","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-02","lu":"2026-04-08"},{"p":"ollama-cloud","id":"ministral-3:3b","n":"ministral-3:3b","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ctx":262144,"out":128000,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2024-10-22","lu":"2026-01-19"},{"p":"ollama-cloud","id":"gemma3:12b","n":"gemma3:12b","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ctx":131072,"out":131072,"mi":"ti","mo":"t","ow":1,"rd":"2024-12-01","lu":"2026-01-19"},{"p":"ollama-cloud","id":"gemma3:4b","n":"gemma3:4b","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ctx":131072,"out":131072,"mi":"ti","mo":"t","ow":1,"rd":"2024-12-01","lu":"2026-01-19"},{"p":"ollama-cloud","id":"qwen3.5:397b","n":"qwen3.5:397b","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-17"},{"p":"ollama-cloud","id":"kimi-k2.6","n":"kimi-k2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"ollama-cloud","id":"qwen3-coder-next","n":"qwen3-coder-next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-02","lu":"2026-02-08"},{"p":"ollama-cloud","id":"qwen3-vl:235b-instruct","n":"qwen3-vl:235b-instruct","f":"qwen","d":"Legacy model retained for compatibility with older integrations","ctx":262144,"out":131072,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-09-22","lu":"2026-01-19"},{"p":"ollama-cloud","id":"mistral-large-3:675b","n":"mistral-large-3:675b","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2026-01-19"},{"p":"ollama-cloud","id":"gemma3:27b","n":"gemma3:27b","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ctx":131072,"out":131072,"mi":"ti","mo":"t","ow":1,"rd":"2025-07-27","lu":"2026-01-19"},{"p":"ollama-cloud","id":"qwen3-vl:235b","n":"qwen3-vl:235b","f":"qwen","d":"Legacy model retained for compatibility with older integrations","ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-22","lu":"2026-01-19"},{"p":"ollama-cloud","id":"nemotron-3-ultra","n":"nemotron-3-ultra","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ctx":262144,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"ollama-cloud","id":"gemini-3-flash-preview","n":"gemini-3-flash-preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ctx":1048576,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-12-17","lu":"2026-04-08"},{"p":"ollama-cloud","id":"gpt-oss:20b","n":"gpt-oss:20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2026-01-19"},{"p":"ollama-cloud","id":"glm-5","n":"glm-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"ollama-cloud","id":"deepseek-v3.2","n":"deepseek-v3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-15","lu":"2026-01-19"},{"p":"ollama-cloud","id":"devstral-2:123b","n":"devstral-2:123b","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-09","lu":"2026-01-19"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/baai/bge-m3","n":"BGE M3","f":"bge","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.012,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/baai/bge-small-en-v1.5","n":"BGE Small EN v1.5","f":"bge","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.02,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/baai/bge-reranker-base","n":"BGE Reranker Base","f":"bge","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0.0031,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-09","lu":"2025-04-09"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/baai/bge-base-en-v1.5","n":"BGE Base EN v1.5","f":"bge","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.067,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/baai/bge-large-en-v1.5","n":"BGE Large EN v1.5","f":"bge","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.2,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/ai4bharat/indictrans2-en-indic-1B","n":"IndicTrans2 EN-Indic 1B","f":"indictrans","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.34,"co":0.34,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/ibm-granite/granite-4.0-h-micro","n":"IBM Granite 4.0 H Micro","f":"granite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.017,"co":0.11,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-10-15","lu":"2025-10-15"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/huggingface/distilbert-sst-2-int8","n":"DistilBERT SST-2 INT8","f":"distilbert","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.026,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-20","lu":"2026-04-20"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/mistral/mistral-7b-instruct-v0.1","n":"Mistral 7B Instruct v0.1","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.11,"co":0.19,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/google/gemma-3-12b-it","n":"Gemma 3 12B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.35,"co":0.56,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-11","lu":"2025-04-11"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/myshell-ai/melotts","n":"MyShell MeloTTS","f":"melotts","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/openai/gpt-oss-120b","n":"GPT OSS 120B","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.35,"co":0.75,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-08-05","lu":"2025-08-05"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/openai/gpt-oss-20b","n":"GPT OSS 20B","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.2,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-08-05","lu":"2025-08-05"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/mistralai/mistral-small-3.1-24b-instruct","n":"Mistral Small 3.1 24B Instruct","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.35,"co":0.56,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-11","lu":"2025-04-11"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/nvidia/nemotron-3-120b-a12b","n":"Nemotron 3 Super 120B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.5,"co":1.5,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/pfnet/plamo-embedding-1b","n":"PLaMo Embedding 1B","f":"plamo","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.019,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/deepgram/aura-2-es","n":"Deepgram Aura 2 (ES)","f":"aura","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/deepgram/aura-2-en","n":"Deepgram Aura 2 (EN)","f":"aura","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/deepgram/nova-3","n":"Deepgram Nova 3","f":"nova","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/zai-org/glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.06,"co":0.4,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/deepseek-ai/deepseek-r1-distill-qwen-32b","n":"DeepSeek R1 Distill Qwen 32B","f":"deepseek-thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.5,"co":4.88,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/qwen/qwen3-embedding-0.6b","n":"Qwen3 Embedding 0.6B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.012,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/qwen/qwen3-30b-a3b-fp8","n":"Qwen3 30B A3B FP8","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.051,"co":0.34,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/qwen/qwen2.5-coder-32b-instruct","n":"Qwen 2.5 Coder 32B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.66,"co":1,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-11","lu":"2025-04-11"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/qwen/qwq-32b","n":"QwQ 32B","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.66,"co":1,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-11","lu":"2025-04-11"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/pipecat-ai/smart-turn-v2","n":"Pipecat Smart Turn v2","f":"smart-turn","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-11-14","lu":"2025-11-14"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.1-8b-instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.28,"co":0.8299999999999998,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/m2m100-1.2b","n":"M2M100 1.2B","f":"m2m","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.34,"co":0.34,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.2-1b-instruct","n":"Llama 3.2 1B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.027,"co":0.2,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.2-11b-vision-instruct","n":"Llama 3.2 11B Vision Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.049,"co":0.68,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout 17B 16E Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.27,"co":0.85,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-16","lu":"2025-04-16"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-guard-3-8b","n":"Llama Guard 3 8B","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.48,"co":0.03,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3-8b-instruct-awq","n":"Llama 3 8B Instruct AWQ","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.12,"co":0.27,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.1-8b-instruct-awq","n":"Llama 3.1 8B Instruct AWQ","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.12,"co":0.27,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.3-70b-instruct-fp8-fast","n":"Llama 3.3 70B Instruct FP8 Fast","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":0.29,"co":2.25,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3-8b-instruct","n":"Llama 3 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.28,"co":0.83,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.1-8b-instruct-fp8","n":"Llama 3.1 8B Instruct FP8","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.15,"co":0.29,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-2-7b-chat-fp16","n":"Llama 2 7B Chat FP16","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.56,"co":6.67,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/meta/llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.051,"co":0.34,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/facebook/bart-large-cnn","n":"BART Large CNN","f":"bart","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-09","lu":"2025-04-09"},{"p":"cloudflare-ai-gateway","id":"workers-ai/@cf/aisingapore/gemma-sea-lion-v4-27b-it","n":"Gemma SEA-LION v4 27B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.35,"co":0.56,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"cloudflare-ai-gateway","id":"openai/o3","n":"o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-3.5-turbo","n":"GPT-3.5-turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"cr":1.25,"ctx":16385,"out":4096,"mi":"t","mo":"t","kn":"2021-09-01","rd":"2023-03-01","lu":"2023-11-06"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"cloudflare-ai-gateway","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.28,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"cloudflare-ai-gateway","id":"openai/o3-pro","n":"o3-pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":20,"co":80,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-06-10","lu":"2025-06-10"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"cloudflare-ai-gateway","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"cloudflare-ai-gateway","id":"openai/o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"cr":0.08,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"cloudflare-ai-gateway","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-3.5-haiku","n":"Claude Haiku 3.5 (latest)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-3.5-sonnet","n":"Claude Sonnet 3.5 v2","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-04-30","rd":"2024-10-22","lu":"2024-10-22"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-opus-4-5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01","rd":"2026-04-16","lu":"2026-04-16"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-3-sonnet","n":"Claude Sonnet 3","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":0.3,"ctx":200000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-03-04","lu":"2024-03-04"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-3-opus","n":"Claude Opus 3","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-02-29","lu":"2024-02-29"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-3-5-haiku","n":"Claude Haiku 3.5 (latest)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-opus-4-1","n":"Claude Opus 4.1 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-opus-4-6","n":"Claude Opus 4.6 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-3-haiku","n":"Claude Haiku 3","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.25,"co":1.25,"cr":0.03,"cw":0.3,"ctx":200000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-03-13","lu":"2024-03-13"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"cloudflare-ai-gateway","id":"anthropic/claude-opus-4","n":"Claude Opus 4 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"moonshotai-cn","id":"kimi-k2.7-code-highspeed","n":"Kimi K2.7 Code HighSpeed","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":1.9,"co":8,"cr":0.38,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"moonshotai-cn","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"moonshotai-cn","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"moonshotai-cn","id":"kimi-k2-turbo-preview","n":"Kimi K2 Turbo","f":"kimi-k2","d":"Fast Kimi model for responsive chat, coding help, and agent loops","ci":2.4,"co":10,"cr":0.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"moonshotai-cn","id":"kimi-k2-0711-preview","n":"Kimi K2 0711","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"cr":0.15,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-07-14","lu":"2025-07-14"},{"p":"moonshotai-cn","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Thinking Kimi model for slower research passes, planning, and hard technical questions","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"moonshotai-cn","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"moonshotai-cn","id":"kimi-k2-thinking-turbo","n":"Kimi K2 Thinking Turbo","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":1.15,"co":8,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"moonshotai-cn","id":"kimi-k2-0905-preview","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"morph","id":"morph-v3-fast","n":"Morph v3 Fast","f":"morph","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.8,"co":1.2,"ctx":16000,"out":16000,"mi":"t","mo":"t","rd":"2024-08-15","lu":"2024-08-15"},{"p":"morph","id":"morph-v3-large","n":"Morph v3 Large","f":"morph","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.9,"co":1.9,"ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2024-08-15","lu":"2024-08-15"},{"p":"morph","id":"auto","n":"Auto","f":"auto","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.85,"co":1.55,"ctx":32000,"out":32000,"mi":"t","mo":"t","rd":"2024-06-01","lu":"2024-06-01"},{"p":"sakana","id":"fugu-ultra-20260615","n":"Fugu Ultra","f":"fugu","d":"Quality-first multi-agent model for hard research, analysis, and competitions","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-15","lu":"2026-06-15"},{"p":"sakana","id":"fugu","n":"Fugu","f":"fugu","d":"Multi-agent model for routing expert agents across complex analytical tasks","ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-15","lu":"2026-06-15"},{"p":"sakana","id":"fugu-ultra","n":"Fugu Ultra","f":"fugu","d":"Quality-first multi-agent model for hard research, analysis, and competitions","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-15","lu":"2026-06-15"},{"p":"deepinfra","id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","n":"Llama 4 Maverick 17B FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.15,"co":0.6,"ctx":1048576,"out":16384,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-05","lu":"2025-04-05"},{"p":"deepinfra","id":"meta-llama/Llama-4-Scout-17B-16E-Instruct","n":"Llama 4 Scout 17B","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.1,"co":0.3,"ctx":327680,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-04-05","lu":"2025-04-05"},{"p":"deepinfra","id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","n":"Llama 3.3 70B Turbo","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":0.1,"co":0.32,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-12-06","lu":"2024-12-06"},{"p":"deepinfra","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.75,"co":3.5,"cr":0.15,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-04","rd":"2026-04-21","lu":"2026-04-21"},{"p":"deepinfra","id":"moonshotai/Kimi-K2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.45,"co":2.25,"cr":0.07,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"deepinfra","id":"moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.74,"co":3.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"deepinfra","id":"google/gemma-4-31B-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.13,"co":0.38,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"deepinfra","id":"google/gemma-4-26B-A4B-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.07,"co":0.34,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"deepinfra","id":"Qwen/Qwen3.6-35B-A3B","n":"Qwen3.6 35B A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.15,"co":0.95,"ctx":262144,"out":81920,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"deepinfra","id":"Qwen/Qwen3.7-Max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.5,"ctx":256000,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"deepinfra","id":"Qwen/Qwen3.6-27B","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.32,"co":3.2,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"deepinfra","id":"Qwen/Qwen3-Next-80B-A3B-Instruct","n":"Qwen3-Next 80B-A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":1.1,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"deepinfra","id":"Qwen/Qwen3-Max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","ci":1.2,"co":6,"cr":0.24,"ctx":256000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"deepinfra","id":"Qwen/Qwen3.5-397B-A17B","n":"Qwen 3.5 397B A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.45,"co":3,"cr":0.22,"ctx":262144,"out":81920,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-01","lu":"2026-04-20"},{"p":"deepinfra","id":"Qwen/Qwen3.5-122B-A10B","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":2.4,"ctx":16384,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"deepinfra","id":"Qwen/Qwen3.5-27B","n":"Qwen3.5 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.6,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"deepinfra","id":"Qwen/Qwen3.5-9B","n":"Qwen3.5 9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.15,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"deepinfra","id":"Qwen/Qwen3-32B","n":"Qwen3 32B","f":"qwen","d":"Dense open Qwen model for self-hosted chat, reasoning, and coding","ci":0.08,"co":0.28,"ctx":40960,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"deepinfra","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo","n":"Qwen3 Coder 480B A35B Instruct Turbo","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.3,"co":1,"cr":0.1,"ctx":262144,"out":66536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"deepinfra","id":"Qwen/Qwen3.5-35B-A3B","n":"Qwen 3.5 35B A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.14,"co":1,"cr":0.05,"ctx":262144,"out":81920,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-01","lu":"2026-04-20"},{"p":"deepinfra","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.039,"co":0.17,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"deepinfra","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.03,"co":0.14,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"deepinfra","id":"XiaomiMiMo/MiMo-V2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.4,"co":2,"cr":0.08,"ctx":262144,"out":16384,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"deepinfra","id":"XiaomiMiMo/MiMo-V2.5-Pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":1,"co":3,"cr":0.2,"ctx":1048576,"out":16384,"mi":"ta","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"deepinfra","id":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning","n":"Nemotron 3 Nano Omni 30B A3B Reasoning","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0.2,"co":0.8,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"deepinfra","id":"nvidia/Nemotron-3-Nano-30B-A3B","n":"Nemotron 3 Nano 30B A3B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.05,"co":0.2,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-15","lu":"2025-12-15"},{"p":"deepinfra","id":"nvidia/Llama-3.3-Nemotron-Super-49B-v1.5","n":"Llama 3.3 Nemotron Super 49B v1.5","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.4,"co":0.4,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-25","lu":"2025-07-25"},{"p":"deepinfra","id":"zai-org/GLM-4.7-Flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.06,"co":0.4,"cr":0.01,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"deepinfra","id":"zai-org/GLM-4.6","n":"GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.43,"co":1.74,"cr":0.08,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"deepinfra","id":"zai-org/GLM-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.08,"cr":0.12,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-02-12","lu":"2026-02-12"},{"p":"deepinfra","id":"zai-org/GLM-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.4,"co":1.75,"cr":0.08,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"deepinfra","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0.93,"co":3,"cr":0.18,"ctx":1048576,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"deepinfra","id":"zai-org/GLM-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.05,"co":3.5,"cr":0.205,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-07","lu":"2026-04-07"},{"p":"deepinfra","id":"deepseek-ai/DeepSeek-R1-0528","n":"DeepSeek-R1-0528","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.15,"cr":0.35,"ctx":163840,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2025-05-28","lu":"2025-05-28"},{"p":"deepinfra","id":"deepseek-ai/DeepSeek-V4-Flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.09,"co":0.18,"cr":0.018,"ctx":1048576,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"deepinfra","id":"deepseek-ai/DeepSeek-V4-Pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.3,"co":2.6,"cr":0.1,"ctx":1048576,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"deepinfra","id":"deepseek-ai/DeepSeek-V3.2","n":"DeepSeek-V3.2","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.26,"co":0.38,"cr":0.13,"ctx":163840,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2025-12-02","lu":"2025-12-02"},{"p":"deepinfra","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.15,"co":1.15,"cr":0.03,"ctx":196608,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-02-12","lu":"2026-02-12"},{"p":"deepinfra","id":"MiniMaxAI/MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.3,"co":1.2,"cr":0.06,"ctx":524288,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"deepinfra","id":"MiniMaxAI/MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.25,"co":1,"cr":0.05,"ctx":196608,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"google-vertex-anthropic","id":"claude-haiku-4-5@20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"google-vertex-anthropic","id":"claude-opus-4@20250514","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"google-vertex-anthropic","id":"claude-opus-4-1@20250805","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"google-vertex-anthropic","id":"claude-opus-4-5@20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-01","lu":"2025-11-01"},{"p":"google-vertex-anthropic","id":"claude-3-5-haiku@20241022","n":"Claude Haiku 3.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"google-vertex-anthropic","id":"claude-sonnet-4@20250514","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"google-vertex-anthropic","id":"claude-opus-4-7@default","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"google-vertex-anthropic","id":"claude-sonnet-4-5@20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"google-vertex-anthropic","id":"claude-sonnet-5@default","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"google-vertex-anthropic","id":"claude-opus-4-6@default","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"google-vertex-anthropic","id":"claude-opus-4-8@default","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"google-vertex-anthropic","id":"claude-sonnet-4-6@default","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"v0","id":"v0-1.0-md","n":"v0-1.0-md","f":"v0","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":3,"co":15,"ctx":128000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"v0","id":"v0-1.5-lg","n":"v0-1.5-lg","f":"v0","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":15,"co":75,"ctx":512000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-06-09","lu":"2025-06-09"},{"p":"v0","id":"v0-1.5-md","n":"v0-1.5-md","f":"v0","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":3,"co":15,"ctx":128000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-06-09","lu":"2025-06-09"},{"p":"azure","id":"codex-mini","n":"Codex Mini","f":"gpt-codex-mini","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.5,"co":6,"cr":0.375,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-05-16","lu":"2025-05-16"},{"p":"azure","id":"phi-3.5-moe-instruct","n":"Phi-3.5-MoE-instruct","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.16,"co":0.64,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"azure","id":"gpt-3.5-turbo-instruct","n":"GPT-3.5 Turbo Instruct","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":4096,"out":4096,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-09-21","lu":"2023-09-21"},{"p":"azure","id":"deepseek-r1-0528","n":"DeepSeek-R1-0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":1.35,"co":5.4,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-05-28","lu":"2025-05-28"},{"p":"azure","id":"deepseek-v4-flash","n":"DeepSeek-V4-Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.19,"co":0.51,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"azure","id":"gpt-5.2-chat","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"azure","id":"o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"azure","id":"deepseek-v3-0324","n":"DeepSeek-V3-0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":1.14,"co":4.56,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2025-03-24","lu":"2025-03-24"},{"p":"azure","id":"phi-3-small-128k-instruct","n":"Phi-3-small-instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.15,"co":0.6,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure","id":"meta-llama-3-8b-instruct","n":"Meta-Llama-3-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.3,"co":0.61,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2023-12","rd":"2024-04-18","lu":"2024-04-18"},{"p":"azure","id":"mistral-small-2503","n":"Mistral Small 3.1","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"ctx":128000,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-09","rd":"2025-03-01","lu":"2025-03-01"},{"p":"azure","id":"text-embedding-3-large","n":"text-embedding-3-large","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.13,"co":0,"ctx":8191,"out":3072,"mi":"t","mo":"t","rd":"2024-01-25","lu":"2024-01-25"},{"p":"azure","id":"o1-mini","n":"o1-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":128000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-09-12","lu":"2024-09-12"},{"p":"azure","id":"phi-3.5-mini-instruct","n":"Phi-3.5-mini-instruct","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.13,"co":0.52,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"azure","id":"mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-07-18","lu":"2024-07-18"},{"p":"azure","id":"cohere-embed-v-4-0","n":"Embed v4","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.12,"co":0,"ctx":128000,"out":1536,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"azure","id":"claude-opus-4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-08-01"},{"p":"azure","id":"gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.13,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure","id":"gpt-5-chat","n":"GPT-5 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.13,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"kn":"2024-10-24","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure","id":"cohere-command-a","n":"Command A","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-03-13","lu":"2025-03-13"},{"p":"azure","id":"llama-3.2-11b-vision-instruct","n":"Llama-3.2-11B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.37,"co":0.37,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"azure","id":"gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"azure","id":"cohere-command-r-08-2024","n":"Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.15,"co":0.6,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"azure","id":"gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"azure","id":"gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":60,"co":120,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-03-14","lu":"2023-03-14"},{"p":"azure","id":"o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"azure","id":"phi-3-medium-128k-instruct","n":"Phi-3-medium-instruct (128k)","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.17,"co":0.68,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure","id":"gpt-4-32k","n":"GPT-4 32K","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":60,"co":120,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-03-14","lu":"2023-03-14"},{"p":"azure","id":"meta-llama-3.1-405b-instruct","n":"Meta-Llama-3.1-405B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":5.33,"co":16,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"azure","id":"cohere-command-r-plus-08-2024","n":"Command R+","f":"command-r","d":"Cohere's RAG workhorse for long-context enterprise search and tool use","ci":2.5,"co":10,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"azure","id":"phi-4-mini","n":"Phi-4-mini","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.075,"co":0.3,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure","id":"gpt-3.5-turbo-1106","n":"GPT-3.5 Turbo 1106","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":2,"ctx":16384,"out":16384,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-11-06","lu":"2023-11-06"},{"p":"azure","id":"llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout 17B 16E Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.2,"co":0.78,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"azure","id":"llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.71,"co":0.71,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"azure","id":"grok-4-20-non-reasoning","n":"Grok 4.20 (Non-Reasoning)","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":262000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2025-09","rd":"2026-04-08","lu":"2026-04-08"},{"p":"azure","id":"gpt-5.1-chat","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"tia","mo":"tia","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure","id":"deepseek-v4-pro","n":"DeepSeek-V4-Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.74,"co":3.48,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"azure","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-12-02"},{"p":"azure","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-11-18","lu":"2025-11-18"},{"p":"azure","id":"claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"azure","id":"deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":1.35,"co":5.4,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-01-20"},{"p":"azure","id":"grok-4-1-fast-non-reasoning","n":"Grok 4.1 Fast (Non-Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2025-06-27","lu":"2025-06-27"},{"p":"azure","id":"gpt-5.4-nano","n":"GPT-5.4 Nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"azure","id":"phi-3-small-8k-instruct","n":"Phi-3-small-instruct (8k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.15,"co":0.6,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure","id":"gpt-5.3-chat","n":"GPT-5.3 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-03","lu":"2026-03-03"},{"p":"azure","id":"gpt-3.5-turbo-0125","n":"GPT-3.5 Turbo 0125","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"ctx":16384,"out":16384,"mi":"t","mo":"t","kn":"2021-08","rd":"2024-01-25","lu":"2024-01-25"},{"p":"azure","id":"cohere-embed-v3-multilingual","n":"Embed v3 Multilingual","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":512,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2023-11-07","lu":"2023-11-07"},{"p":"azure","id":"gpt-3.5-turbo-0613","n":"GPT-3.5 Turbo 0613","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":4,"ctx":16384,"out":16384,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-06-13","lu":"2023-06-13"},{"p":"azure","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"tia","mo":"tia","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure","id":"gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"azure","id":"phi-4-reasoning","n":"Phi-4-reasoning","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.125,"co":0.5,"ctx":32000,"out":4096,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-31","rd":"2026-05-28","lu":"2026-05-28"},{"p":"azure","id":"mistral-medium-2505","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2025-05","rd":"2025-05-07","lu":"2025-05-07"},{"p":"azure","id":"meta-llama-3.1-70b-instruct","n":"Meta-Llama-3.1-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.68,"co":3.54,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"azure","id":"text-embedding-ada-002","n":"text-embedding-ada-002","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2022-12-15","lu":"2022-12-15"},{"p":"azure","id":"o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"azure","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"azure","id":"gpt-4-turbo-vision","n":"GPT-4 Turbo Vision","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"azure","id":"mistral-large-2411","n":"Mistral Large 24.11","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2024-09","rd":"2024-11-01","lu":"2024-11-01"},{"p":"azure","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-24","lu":"2026-02-24"},{"p":"azure","id":"text-embedding-3-small","n":"text-embedding-3-small","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.02,"co":0,"ctx":8191,"out":1536,"mi":"t","mo":"t","rd":"2024-01-25","lu":"2024-01-25"},{"p":"azure","id":"deepseek-v3.2-speciale","n":"DeepSeek-V3.2-Speciale","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.58,"co":1.68,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"azure","id":"gpt-3.5-turbo-0301","n":"GPT-3.5 Turbo 0301","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":4096,"out":4096,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-03-01","lu":"2023-03-01"},{"p":"azure","id":"meta-llama-3.1-8b-instruct","n":"Meta-Llama-3.1-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.3,"co":0.61,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"azure","id":"claude-opus-4-1","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-18","lu":"2025-11-18"},{"p":"azure","id":"gpt-5.1-codex-mini","n":"GPT-5.1 Codex Mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-06","lu":"2026-02-06"},{"p":"azure","id":"phi-4","n":"Phi-4","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.125,"co":0.5,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure","id":"grok-4-fast-reasoning","n":"Grok 4 Fast (Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-07","rd":"2025-09-19","lu":"2025-09-19"},{"p":"azure","id":"gpt-image-1.5","n":"GPT-Image-1.5","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":32,"cr":1.25,"ctx":0,"out":0,"mi":"ti","mo":"ti","rd":"2025-11-25","lu":"2025-11-25"},{"p":"azure","id":"gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"azure","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"azure","id":"o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"azure","id":"deepseek-v3.1","n":"DeepSeek-V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.56,"co":1.68,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-08-21","lu":"2025-08-21"},{"p":"azure","id":"llama-3.2-90b-vision-instruct","n":"Llama-3.2-90B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":2.04,"co":2.04,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"azure","id":"claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-31","rd":"2025-11-18","lu":"2025-11-18"},{"p":"azure","id":"ministral-3b","n":"Ministral 3B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.04,"co":0.04,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-03","rd":"2024-10-22","lu":"2024-10-22"},{"p":"azure","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"azure","id":"gpt-5.4-mini","n":"GPT-5.4 Mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"azure","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-22","lu":"2026-04-22"},{"p":"azure","id":"gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"azure","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"azure","id":"grok-4-1-fast-reasoning","n":"Grok 4.1 Fast (Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":128000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-06-27","lu":"2025-06-27"},{"p":"azure","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.03,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure","id":"gpt-image-1","n":"GPT-Image-1","f":"gpt-image","d":"OpenAI image model for production generation, edits, and brand-safe visual workflows","ci":5,"co":40,"cr":1.25,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-04-24","lu":"2025-04-24"},{"p":"azure","id":"gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"azure","id":"gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"azure","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.01,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure","id":"phi-3-medium-4k-instruct","n":"Phi-3-medium-instruct (4k)","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.17,"co":0.68,"ctx":4096,"out":1024,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure","id":"gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"azure","id":"cohere-embed-v3-english","n":"Embed v3 English","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":512,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2023-11-07","lu":"2023-11-07"},{"p":"azure","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"azure","id":"codestral-2501","n":"Codestral 25.01","f":"codestral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.9,"ctx":256000,"out":256000,"mi":"t","mo":"t","tc":1,"kn":"2024-03","rd":"2025-01-01","lu":"2025-01-01"},{"p":"azure","id":"phi-4-reasoning-plus","n":"Phi-4-reasoning-plus","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.125,"co":0.5,"ctx":32000,"out":4096,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure","id":"phi-4-multimodal","n":"Phi-4-multimodal","f":"phi","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.08,"co":0.32,"ctx":128000,"out":4096,"mi":"tia","mo":"t","ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure","id":"gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"azure","id":"model-router","n":"Model Router","f":"model-router","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.14,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-05-19","lu":"2025-11-18"},{"p":"azure","id":"grok-4-20-reasoning","n":"Grok 4.20 (Reasoning)","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":262000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-09","rd":"2026-04-08","lu":"2026-04-08"},{"p":"azure","id":"gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"azure","id":"deepseek-v3.2","n":"DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.58,"co":1.68,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"azure","id":"llama-4-maverick-17b-128e-instruct-fp8","n":"Llama 4 Maverick 17B 128E Instruct FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.25,"co":1,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"azure","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-14","lu":"2026-01-14"},{"p":"azure","id":"phi-4-mini-reasoning","n":"Phi-4-mini-reasoning","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.075,"co":0.3,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure","id":"phi-3-mini-128k-instruct","n":"Phi-3-mini-instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.13,"co":0.52,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure","id":"gpt-image-2","n":"GPT-Image-2","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5,"co":30,"cr":1.25,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2026-04-21","lu":"2026-04-21"},{"p":"azure","id":"phi-3-mini-4k-instruct","n":"Phi-3-mini-instruct (4k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.13,"co":0.52,"ctx":4096,"out":1024,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure","id":"meta-llama-3-70b-instruct","n":"Meta-Llama-3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.68,"co":3.54,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2023-12","rd":"2024-04-18","lu":"2024-04-18"},{"p":"azure","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.125,"ctx":272000,"out":128000,"mi":"tia","mo":"tia","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-24","lu":"2026-04-24"},{"p":"cerebras","id":"gemma-4-31b","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.99,"co":1.49,"ctx":131072,"out":40960,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-07-01"},{"p":"cerebras","id":"gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.35,"co":0.75,"ctx":131072,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2026-06-10"},{"p":"cerebras","id":"zai-glm-4.7","n":"Z.AI GLM-4.7","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":2.25,"co":2.75,"cr":0,"cw":0,"ctx":131072,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-07","lu":"2026-06-10"},{"p":"zai-coding-plan","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"zai-coding-plan","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"zai-coding-plan","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"zai-coding-plan","id":"glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"zai-coding-plan","id":"glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zai-coding-plan","id":"glm-5-turbo","n":"GLM-5-Turbo","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"nvidia","id":"baai/bge-m3","n":"BGE M3","f":"bge","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0,"co":0,"ctx":8192,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2024-01-30","lu":"2026-04-30"},{"p":"nvidia","id":"moonshotai/kimi-k2-instruct-0905","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"nvidia","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"nvidia","id":"minimaxai/minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0,"co":0,"ctx":1000000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"nvidia","id":"minimaxai/minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-04-11"},{"p":"nvidia","id":"stepfun-ai/step-3.7-flash","n":"Step 3.7 Flash","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0,"co":0,"ctx":256000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"nvidia","id":"stepfun-ai/step-3.5-flash","n":"Step 3.5 Flash","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0,"co":0,"ctx":256000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-02","lu":"2026-02-02"},{"p":"nvidia","id":"google/gemma-3n-e4b-it","n":"Gemma 3n E4b It","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-06","rd":"2025-06-03","lu":"2025-06-03"},{"p":"nvidia","id":"google/gemma-3n-e2b-it","n":"Gemma 3n E2b It","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-06","rd":"2025-06-12","lu":"2025-06-12"},{"p":"nvidia","id":"google/google-paligemma","n":"paligemma","d":"Gemini multimodal model for text, image, audio, video, and document tasks","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2024-05-14","lu":"2024-08-26"},{"p":"nvidia","id":"google/gemma-4-31b-it","n":"Gemma-4-31B-IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":256000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-02","lu":"2026-04-02"},{"p":"nvidia","id":"google/gemma-2-2b-it","n":"Gemma 2 2b It","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-16","lu":"2024-07-16"},{"p":"nvidia","id":"microsoft/phi-4-mini-instruct","n":"Phi-4-Mini","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0,"co":0,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2024-12-01","lu":"2025-09-05"},{"p":"nvidia","id":"microsoft/phi-4-multimodal-instruct","n":"Phi 4 Multimodal","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nvidia","id":"z-ai/glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"nvidia","id":"openai/gpt-oss-120b","n":"GPT-OSS-120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-04","lu":"2025-08-14"},{"p":"nvidia","id":"openai/whisper-large-v3","n":"Whisper Large v3","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"a","mo":"t","ow":1,"kn":"2023-09","rd":"2023-09-01","lu":"2025-09-05"},{"p":"nvidia","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"nvidia","id":"bytedance/seed-oss-36b-instruct","n":"ByteDance-Seed/Seed-OSS-36B-Instruct","f":"seed","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0,"co":0,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-04","lu":"2025-11-25"},{"p":"nvidia","id":"mistralai/mistral-7b-instruct-v03","n":"Mistral-7B-Instruct-v0.3","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":65536,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-01","lu":"2025-04-01"},{"p":"nvidia","id":"mistralai/magistral-small-2506","n":"Magistral Small 2506","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":0,"co":0,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nvidia","id":"mistralai/mixtral-8x7b-instruct","n":"Mistral: Mixtral 8x7B Instruct","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":32768,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2023-12-10","lu":"2026-03-15"},{"p":"nvidia","id":"mistralai/mistral-medium-3-instruct","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nvidia","id":"mistralai/mistral-small-4-119b-2603","n":"mistral-small-4-119b-2603","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"nvidia","id":"mistralai/mistral-nemotron","n":"mistral-nemotron","f":"nemotron","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-06-11","lu":"2025-06-12"},{"p":"nvidia","id":"mistralai/mistral-large-3-675b-instruct-2512","n":"Mistral Large 3 675B Instruct 2512","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-01","rd":"2025-12-02","lu":"2025-12-02"},{"p":"nvidia","id":"mistralai/mixtral-8x22b-instruct","n":"Mistral: Mixtral 8x22B Instruct","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":65536,"out":13108,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-04-17","lu":"2024-04-17"},{"p":"nvidia","id":"nvidia/cosmos-transfer1-7b","n":"cosmos-transfer1-7b","d":"Video model for prompt-guided generation, editing, and motion workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"tiv","mo":"v","ow":1,"rd":"2025-06-13","lu":"2025-06-30"},{"p":"nvidia","id":"nvidia/cosmos-transfer2_5-2b","n":"cosmos-transfer2.5-2b","d":"Video model for prompt-guided generation, editing, and motion workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"tiv","mo":"v","ow":1,"rd":"2026-02-26","lu":"2026-02-26"},{"p":"nvidia","id":"nvidia/llama-nemotron-embed-vl-1b-v2","n":"llama-nemotron-embed-vl-1b-v2","f":"nemotron","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0,"co":0,"ctx":32768,"out":2048,"mi":"ti","mo":"t","ow":1,"rd":"2026-02-10","lu":"2026-02-10"},{"p":"nvidia","id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning","n":"Nemotron 3 Nano Omni","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0,"co":0,"ctx":256000,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"nvidia","id":"nvidia/magpie-tts-zeroshot","n":"magpie-tts-zeroshot","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":0,"out":4096,"mi":"ta","mo":"a","ow":1,"rd":"2025-05-22","lu":"2025-06-12"},{"p":"nvidia","id":"nvidia/nvidia-nemotron-nano-9b-v2","n":"nvidia-nemotron-nano-9b-v2","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0,"co":0,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-09","rd":"2025-08-18","lu":"2025-08-18"},{"p":"nvidia","id":"nvidia/synthetic-video-detector","n":"synthetic-video-detector","d":"Video model for prompt-guided generation, editing, and motion workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"v","mo":"t","ow":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"nvidia","id":"nvidia/nemotron-content-safety-reasoning-4b","n":"nemotron-content-safety-reasoning-4b","f":"nemotron","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2026-01-22","lu":"2026-01-22"},{"p":"nvidia","id":"nvidia/nv-embed-v1","n":"nv-embed-v1","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0,"co":0,"ctx":32768,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2024-06-07","lu":"2025-07-22"},{"p":"nvidia","id":"nvidia/usdcode","n":"usdcode","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2026-01-01","lu":"2026-01-01"},{"p":"nvidia","id":"nvidia/riva-translate-4b-instruct-v1_1","n":"riva-translate-4b-instruct-v1_1","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2025-12-12","lu":"2025-12-12"},{"p":"nvidia","id":"nvidia/sparsedrive","n":"sparsedrive","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"v","mo":"t","ow":1,"rd":"2025-03-18","lu":"2025-07-20"},{"p":"nvidia","id":"nvidia/rerank-qa-mistral-4b","n":"rerank-qa-mistral-4b","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2024-03-17","lu":"2025-01-17"},{"p":"nvidia","id":"nvidia/streampetr","n":"streampetr","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"v","mo":"t","ow":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"nvidia","id":"nvidia/active-speaker-detection","n":"Active Speaker Detection","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"v","mo":"t","ow":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"nvidia","id":"nvidia/llama-3_1-nemotron-safety-guard-8b-v3","n":"llama-3.1-nemotron-safety-guard-8b-v3","f":"nemotron","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2025-10-28","lu":"2025-10-28"},{"p":"nvidia","id":"nvidia/llama-3_2-nemoretriever-300m-embed-v1","n":"llama-3_2-nemoretriever-300m-embed-v1","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0,"co":0,"ctx":32768,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2025-07-24","lu":"2025-07-24"},{"p":"nvidia","id":"nvidia/nemotron-voicechat","n":"nemotron-voicechat","f":"nemotron","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ta","mo":"t","tc":1,"ow":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"nvidia","id":"nvidia/nv-embedcode-7b-v1","n":"nv-embedcode-7b-v1","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0,"co":0,"ctx":32768,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2025-03-17","lu":"2025-05-29"},{"p":"nvidia","id":"nvidia/nemotron-3-ultra-550b-a55b","n":"Nemotron 3 Ultra 550B A55B","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.5,"co":2.5,"cr":0.15,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"nvidia","id":"nvidia/nemotron-3-nano-30b-a3b","n":"nemotron-3-nano-30b-a3b","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0,"co":0,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-09","rd":"2024-12","lu":"2024-12"},{"p":"nvidia","id":"nvidia/cosmos-predict1-5b","n":"cosmos-predict1-5b","d":"Video model for prompt-guided generation, editing, and motion workflows","ci":0,"co":0,"ctx":0,"out":4096,"mi":"tiv","mo":"v","ow":1,"rd":"2025-03-18","lu":"2025-03-18"},{"p":"nvidia","id":"nvidia/bevformer","n":"bevformer","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"v","mo":"t","ow":1,"rd":"2025-03-18","lu":"2025-07-20"},{"p":"nvidia","id":"nvidia/studiovoice","n":"studiovoice","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-10-03","lu":"2025-06-13"},{"p":"nvidia","id":"nvidia/gliner-pii","n":"gliner-pii","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2026-03-03","lu":"2026-03-03"},{"p":"nvidia","id":"nvidia/nemotron-mini-4b-instruct","n":"nemotron-mini-4b-instruct","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-08-21","lu":"2024-08-26"},{"p":"nvidia","id":"nvidia/llama-nemotron-rerank-vl-1b-v2","n":"llama-nemotron-rerank-vl-1b-v2","f":"nemotron","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","ow":1,"rd":"2026-03-31","lu":"2026-03-31"},{"p":"nvidia","id":"nvidia/nemotron-3-super-120b-a12b","n":"Nemotron 3 Super","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.2,"co":0.8,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-04","rd":"2026-03-11","lu":"2026-03-11"},{"p":"nvidia","id":"nvidia/usdvalidate","n":"usdvalidate","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0,"co":0,"ctx":0,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2024-07-24","lu":"2025-01-08"},{"p":"nvidia","id":"nvidia/nemotron-3-content-safety","n":"nemotron-3-content-safety","f":"nemotron","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"nvidia","id":"abacusai/dracarys-llama-3_1-70b-instruct","n":"dracarys-llama-3.1-70b-instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09-11","lu":"2025-05-22"},{"p":"nvidia","id":"deepseek-ai/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1048576,"out":393216,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"nvidia","id":"deepseek-ai/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1048576,"out":393216,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"nvidia","id":"qwen/qwen3-next-80b-a3b-instruct","n":"Qwen3-Next-80B-A3B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2024-12-01","lu":"2025-09-05"},{"p":"nvidia","id":"qwen/qwen-image-edit","n":"Qwen Image Edit","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-08-19","lu":"2025-08-19"},{"p":"nvidia","id":"qwen/qwen-image","n":"Qwen Image","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2025-08-07","lu":"2025-08-07"},{"p":"nvidia","id":"qwen/qwen3-coder-480b-a35b-instruct","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":262144,"out":66536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"nvidia","id":"qwen/qwen2.5-coder-32b-instruct","n":"Qwen2.5 Coder 32b Instruct","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-06","lu":"2024-11-06"},{"p":"nvidia","id":"qwen/qwen3.5-397b-a17b","n":"Qwen3.5-397B-A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"ctx":262144,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-02-16","lu":"2026-02-16"},{"p":"nvidia","id":"qwen/qwen3.5-122b-a10b","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"nvidia","id":"sarvamai/sarvam-m","n":"sarvam-m","d":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-25","lu":"2025-07-25"},{"p":"nvidia","id":"meta/llama-3.1-8b-instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":16000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"nvidia","id":"meta/llama-3.1-70b-instruct","n":"Llama 3.1 70b Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-16","lu":"2024-07-16"},{"p":"nvidia","id":"meta/llama-3.2-1b-instruct","n":"Llama 3.2 1b Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-18","lu":"2024-09-18"},{"p":"nvidia","id":"meta/llama-3.2-11b-vision-instruct","n":"Llama 3.2 11b Vision Instruct","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-18","lu":"2024-09-18"},{"p":"nvidia","id":"meta/llama-3.3-70b-instruct","n":"Llama 3.3 70b Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-26","lu":"2024-11-26"},{"p":"nvidia","id":"meta/llama-guard-4-12b","n":"Llama Guard 4 12B","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-05","lu":"2026-04-30"},{"p":"nvidia","id":"meta/esmfold","n":"esmfold","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-03-15","lu":"2025-06-12"},{"p":"nvidia","id":"meta/esm2-650m","n":"esm2-650m","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-08-29","lu":"2025-03-10"},{"p":"nvidia","id":"meta/llama-3.2-90b-vision-instruct","n":"Llama-3.2-90B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"nvidia","id":"meta/llama-4-maverick-17b-128e-instruct","n":"Llama 4 Maverick 17b 128e Instruct","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-02","rd":"2025-04-01","lu":"2025-04-01"},{"p":"nvidia","id":"meta/llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":32768,"out":32000,"mi":"t","mo":"t","ow":1,"rd":"2024-09-18","lu":"2024-09-18"},{"p":"nvidia","id":"upstage/solar-10_7b-instruct","n":"solar-10.7b-instruct","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-06-05","lu":"2025-04-10"},{"p":"nvidia","id":"black-forest-labs/flux_1-schnell","n":"FLUX.1-schnell","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":77,"out":0,"mi":"t","mo":"i","ow":1,"kn":"2024-07","rd":"2024-08-01","lu":"2026-02-04"},{"p":"nvidia","id":"black-forest-labs/flux_1-kontext-dev","n":"FLUX.1-Kontext-dev","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":40960,"out":40960,"mi":"ti","mo":"i","ow":1,"rd":"2025-08-12","lu":"2025-08-12"},{"p":"nvidia","id":"black-forest-labs/flux.1-dev","n":"FLUX.1-dev","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":4096,"out":0,"mi":"t","mo":"i","kn":"2024-08","rd":"2024-08-01","lu":"2025-09-05"},{"p":"nvidia","id":"black-forest-labs/flux_2-klein-4b","n":"FLUX.2 Klein 4B","f":"flux","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":40960,"out":40960,"mi":"it","mo":"i","ow":1,"kn":"2025-06","rd":"2026-01-14","lu":"2026-01-31"},{"p":"evroc","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":1.4375,"co":5.75,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"evroc","id":"google/gemma-4-26B-A4B-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.144,"co":0.575,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"evroc","id":"Qwen/Qwen3-Embedding-8B","n":"Qwen3 Embedding 8B","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.115,"co":0.115,"ctx":40960,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2025-07-30","lu":"2025-07-30"},{"p":"evroc","id":"Qwen/Qwen3-Reranker-4B","n":"Qwen3 Reranker 4B","f":"qwen","d":"Reranking model for improving retrieval quality in search and recommendation systems","ci":0.0575,"co":0,"ctx":32000,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2025-07-30","lu":"2025-07-30"},{"p":"evroc","id":"Qwen/Qwen3.6-35B-A3B-FP8","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.345,"co":1.38,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"evroc","id":"Qwen/Qwen3-VL-30B-A3B-Instruct","n":"Qwen3 VL 30B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.23,"co":0.92,"ctx":100000,"out":100000,"mi":"tiv","mo":"t","tc":1,"ow":1,"rd":"2025-07-30","lu":"2025-07-30"},{"p":"evroc","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.23,"co":0.92,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"evroc","id":"openai/whisper-large-v3-turbo","n":"Whisper Large v3 Turbo","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0.0023,"co":0.0023,"ctx":448,"out":448,"mi":"a","mo":"t","ow":1,"rd":"2024-10-01","lu":"2024-10-01"},{"p":"evroc","id":"openai/whisper-large-v3","n":"Whisper 3 Large","f":"whisper","d":"Open Whisper checkpoint for robust multilingual transcription and captioning","ci":0.0023,"co":0.0023,"ctx":448,"out":4096,"mi":"a","mo":"t","ow":1,"rd":"2024-10-01","lu":"2024-10-01"},{"p":"evroc","id":"mistralai/Mistral-Medium-3.5-128B","n":"Mistral Medium 3.5","f":"mistral-medium","d":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","ci":1.725,"co":6.9,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"evroc","id":"mistralai/Voxtral-Small-24B-2507","n":"Voxtral Small 24B","f":"voxtral","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.0023,"co":0.0023,"ctx":32000,"out":32000,"mi":"at","mo":"t","ow":1,"rd":"2025-03-01","lu":"2025-03-01"},{"p":"evroc","id":"nvidia/Llama-3.3-70B-Instruct-FP8","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":1.15,"co":1.15,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"evroc","id":"evroc/roc","n":"roc","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":2.875,"co":11.516,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2026-01","rd":"2026-06-06","lu":"2026-06-06"},{"p":"evroc","id":"KBLab/kb-whisper-large","n":"KB Whisper","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0.0023,"co":0.0023,"ctx":448,"out":448,"mi":"a","mo":"t","ow":1,"rd":"2024-10-01","lu":"2024-10-01"},{"p":"evroc","id":"intfloat/multilingual-e5-large-instruct","n":"E5 Multi-Lingual Large Embeddings 0.6B","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.114,"co":0.114,"ctx":512,"out":512,"mi":"t","mo":"t","ow":1,"rd":"2024-06-01","lu":"2024-06-01"},{"p":"xiaomi","id":"mimo-v2.5-pro-ultraspeed","n":"MiMo-V2.5-Pro-UltraSpeed","f":"mimo","d":"MiMo pro model for strong multimodal reasoning and agent execution","ci":1.305,"co":2.61,"cr":0.0108,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-06-08","lu":"2026-06-09"},{"p":"xiaomi","id":"mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-06-24"},{"p":"xiaomi","id":"mimo-v2-omni","n":"MiMo-V2-Omni","f":"mimo","d":"Legacy model retained for compatibility with older integrations","ci":0.14,"co":0.28,"cr":0.0028,"ctx":262144,"out":131072,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-06-24"},{"p":"xiaomi","id":"mimo-v2-flash","n":"MiMo-V2-Flash","f":"mimo","d":"Legacy model retained for compatibility with older integrations","ci":0.14,"co":0.28,"cr":0.0028,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12-01","rd":"2025-12-16","lu":"2026-06-24"},{"p":"xiaomi","id":"mimo-v2-pro","n":"MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-06-24"},{"p":"xiaomi","id":"mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-06-24"},{"p":"inception","id":"mercury-edit-2","n":"Mercury Edit 2","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.25,"co":0.75,"cr":0.025,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"rd":"2026-03-30","lu":"2026-03-30"},{"p":"inception","id":"mercury-2","n":"Mercury 2","f":"mercury","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.25,"co":0.75,"cr":0.025,"ctx":128000,"out":50000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-02-24","lu":"2026-02-24"},{"p":"anthropic","id":"claude-opus-4-5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"anthropic","id":"claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"anthropic","id":"claude-opus-4-1-20250805","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"anthropic","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"anthropic","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-14","lu":"2026-04-16"},{"p":"anthropic","id":"claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-29","lu":"2026-06-30"},{"p":"anthropic","id":"claude-opus-4-5-20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-01"},{"p":"anthropic","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"anthropic","id":"claude-opus-4-1","n":"Claude Opus 4.1 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"anthropic","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-07","lu":"2026-06-09"},{"p":"anthropic","id":"claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"anthropic","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-04","lu":"2026-03-13"},{"p":"anthropic","id":"claude-sonnet-4-5-20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"anthropic","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"tencent-coding-plan","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"tencent-coding-plan","id":"kimi-k2.5","n":"Kimi-K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"tencent-coding-plan","id":"hunyuan-turbos","n":"Hunyuan-TurboS","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2026-03-08","lu":"2026-03-08"},{"p":"tencent-coding-plan","id":"hunyuan-t1","n":"Hunyuan-T1","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-08","lu":"2026-03-08"},{"p":"tencent-coding-plan","id":"tc-code-latest","n":"Auto","f":"auto","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2026-03-08","lu":"2026-03-08"},{"p":"tencent-coding-plan","id":"glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"tencent-coding-plan","id":"hunyuan-2.0-instruct","n":"Tencent HY 2.0 Instruct","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2026-03-08","lu":"2026-03-08"},{"p":"tencent-coding-plan","id":"hunyuan-2.0-thinking","n":"Tencent HY 2.0 Think","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-08","lu":"2026-03-08"},{"p":"freemodel","id":"claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"freemodel","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"freemodel","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"freemodel","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"cw":1.75,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"freemodel","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"freemodel","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"cw":2.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"freemodel","id":"gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"cw":0.75,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"freemodel","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"freemodel","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"freemodel","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"cw":5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"sap-ai-core","id":"anthropic--claude-4.6-sonnet","n":"anthropic--claude-4.6-sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08","rd":"2026-02-17","lu":"2026-03-13"},{"p":"sap-ai-core","id":"anthropic--claude-3-sonnet","n":"anthropic--claude-3-sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-03-04","lu":"2024-03-04"},{"p":"sap-ai-core","id":"anthropic--claude-4-sonnet","n":"anthropic--claude-4-sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"sap-ai-core","id":"gemini-2.5-pro","n":"gemini-2.5-pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-25","lu":"2025-06-05"},{"p":"sap-ai-core","id":"gpt-5","n":"gpt-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"sap-ai-core","id":"gemini-2.5-flash","n":"gemini-2.5-flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-04-17","lu":"2025-06-05"},{"p":"sap-ai-core","id":"anthropic--claude-4.5-haiku","n":"anthropic--claude-4.5-haiku","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"sap-ai-core","id":"anthropic--claude-3-haiku","n":"anthropic--claude-3-haiku","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.25,"co":1.25,"cr":0.03,"cw":0.3,"ctx":200000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-03-13","lu":"2024-03-13"},{"p":"sap-ai-core","id":"anthropic--claude-4-opus","n":"anthropic--claude-4-opus","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"sap-ai-core","id":"anthropic--claude-4.5-sonnet","n":"anthropic--claude-4.5-sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"sap-ai-core","id":"anthropic--claude-3.5-sonnet","n":"anthropic--claude-3.5-sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-04-30","rd":"2024-10-22","lu":"2024-10-22"},{"p":"sap-ai-core","id":"anthropic--claude-4.6-opus","n":"anthropic--claude-4.6-opus","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05","rd":"2026-02-05","lu":"2026-03-13"},{"p":"sap-ai-core","id":"gemini-2.5-flash-lite","n":"gemini-2.5-flash-lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"sap-ai-core","id":"anthropic--claude-3.7-sonnet","n":"anthropic--claude-3.7-sonnet","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-10-31","rd":"2025-02-24","lu":"2025-02-24"},{"p":"sap-ai-core","id":"sonar","n":"sonar","f":"sonar","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":1,"co":1,"ctx":128000,"out":4096,"mi":"t","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"sap-ai-core","id":"sonar-pro","n":"sonar-pro","f":"sonar-pro","d":"Advanced Sonar search model for deeper research and cited synthesis","ci":3,"co":15,"ctx":200000,"out":8192,"mi":"ti","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"sap-ai-core","id":"gpt-5.4","n":"gpt-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"sap-ai-core","id":"gpt-4.1","n":"gpt-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"sap-ai-core","id":"anthropic--claude-3-opus","n":"anthropic--claude-3-opus","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-02-29","lu":"2024-02-29"},{"p":"sap-ai-core","id":"gpt-5-mini","n":"gpt-5-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"sap-ai-core","id":"gpt-4.1-mini","n":"gpt-4.1-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"sap-ai-core","id":"sonar-deep-research","n":"sonar-deep-research","f":"sonar-deep-research","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":2,"co":8,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"kn":"2025-01","rd":"2025-02-01","lu":"2025-09-01"},{"p":"sap-ai-core","id":"gpt-5-nano","n":"gpt-5-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"sap-ai-core","id":"anthropic--claude-4.5-opus","n":"anthropic--claude-4.5-opus","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05","rd":"2025-11-24","lu":"2025-11-24"},{"p":"sap-ai-core","id":"anthropic--claude-4.7-opus","n":"anthropic--claude-4.7-opus","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"sap-ai-core","id":"gpt-5.5","n":"gpt-5.5","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"opencode","id":"ring-2.6-1t-free","n":"Ring 2.6 1T Free","f":"ring-1t-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"ctx":262000,"out":66000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-05-08","lu":"2026-05-08"},{"p":"opencode","id":"mimo-v2-pro-free","n":"MiMo V2 Pro Free","f":"mimo-pro-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":1048576,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"opencode","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"opencode","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Legacy model retained for compatibility with older integrations","ci":0.6,"co":2.2,"cr":0.1,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"opencode","id":"mimo-v2.5-free","n":"MiMo V2.5 Free","f":"mimo-v2.5-free","d":"MiMo omni model for text, image, video, audio, and agents","ci":0,"co":0,"cr":0,"ctx":200000,"out":32000,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode","id":"kimi-k2","n":"Kimi K2","f":"kimi-k2","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2.5,"cr":0.4,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"opencode","id":"minimax-m2.1","n":"MiniMax-M2.1","f":"minimax","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":1.2,"cr":0.1,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-12-23","lu":"2025-12-23"},{"p":"opencode","id":"nemotron-3-ultra-free","n":"Nemotron 3 Ultra Free","f":"nemotron-free","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0,"co":0,"cr":0,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-02","rd":"2026-06-04","lu":"2026-06-04"},{"p":"opencode","id":"glm-4.7-free","n":"GLM-4.7 Free","f":"glm-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"opencode","id":"gemini-3-flash","n":"Gemini 3 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"opencode","id":"deepseek-v4-flash-free","n":"DeepSeek V4 Flash Free","f":"deepseek-flash-free","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0,"co":0,"cr":0,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode","id":"claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"opencode","id":"claude-opus-4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"opencode","id":"gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.07,"co":8.5,"cr":0.107,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"opencode","id":"gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"opencode","id":"minimax-m3-free","n":"MiniMax-M3 Free","f":"minimax-m3-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":200000,"out":32000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-05-31","lu":"2026-05-31"},{"p":"opencode","id":"gemini-3-pro","n":"Gemini 3 Pro","f":"gemini-pro","d":"Legacy model retained for compatibility with older integrations","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"opencode","id":"kimi-k2.5-free","n":"Kimi K2.5 Free","f":"kimi-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2026-01-27","lu":"2026-01-27"},{"p":"opencode","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"opencode","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cr":0.26,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-07","lu":"2026-04-07"},{"p":"opencode","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.74,"co":3.84,"cr":0.145,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode","id":"glm-4.6","n":"GLM-4.6","f":"glm","d":"Legacy model retained for compatibility with older integrations","ci":0.6,"co":2.2,"cr":0.1,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"opencode","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Legacy model retained for compatibility with older integrations","ci":0.4,"co":2.5,"cr":0.4,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2025-09-05","lu":"2025-09-05"},{"p":"opencode","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"opencode","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"opencode","id":"claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"opencode","id":"gpt-5.4-nano","n":"GPT-5.4 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"opencode","id":"qwen3-coder","n":"Qwen3 Coder","f":"qwen","d":"Legacy model retained for compatibility with older integrations","ci":0.45,"co":1.8,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"opencode","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"opencode","id":"hy3-preview-free","n":"Hy3 preview Free","f":"hy3-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-04-20","lu":"2026-04-20"},{"p":"opencode","id":"minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.3,"co":1.2,"cr":0.06,"ctx":512000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"opencode","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.07,"co":8.5,"cr":0.107,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"opencode","id":"gpt-5.3-codex-spark","n":"GPT-5.3 Codex Spark","f":"gpt-codex-spark","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-12","lu":"2026-02-12"},{"p":"opencode","id":"gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"opencode","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"opencode","id":"ling-2.6-flash-free","n":"Ling 2.6 Flash Free","f":"ling-flash-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"ctx":262100,"out":32800,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-06","rd":"2026-04-21","lu":"2026-04-21"},{"p":"opencode","id":"qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen3.5","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.2,"co":1.2,"cr":0.02,"cw":0.25,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"opencode","id":"claude-3-5-haiku","n":"Claude Haiku 3.5","f":"claude-haiku","d":"Legacy model retained for compatibility with older integrations","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"opencode","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"opencode","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-24","lu":"2026-02-24"},{"p":"opencode","id":"minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-03-18","lu":"2026-03-18"},{"p":"opencode","id":"north-mini-code-free","n":"North Mini Code Free","f":"north-free","d":"Cohere coding model for practical software engineering and agentic edits","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09-23","rd":"2026-06-09","lu":"2026-06-09"},{"p":"opencode","id":"grok-code","n":"Grok Code Fast 1","f":"grok","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"cw":0,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-20","lu":"2025-08-20"},{"p":"opencode","id":"claude-opus-4-1","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"opencode","id":"gpt-5.1-codex-mini","n":"GPT-5.1 Codex Mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"opencode","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.08,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2026-01-27","lu":"2026-01-27"},{"p":"opencode","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"opencode","id":"claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"opencode","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"opencode","id":"gpt-5.4-mini","n":"GPT-5.4 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"opencode","id":"minimax-m2.5-free","n":"MiniMax-M2.5 Free","f":"minimax-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"opencode","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2026-04-21","lu":"2026-04-21"},{"p":"opencode","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"opencode","id":"minimax-m2.1-free","n":"MiniMax-M2.1 Free","f":"minimax-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-12-23","lu":"2025-12-23"},{"p":"opencode","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"opencode","id":"gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"cr":30,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"opencode","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"opencode","id":"mimo-v2-flash-free","n":"MiMo V2 Flash Free","f":"mimo-flash-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-12-16","lu":"2025-12-16"},{"p":"opencode","id":"gemini-3.1-pro","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"opencode","id":"trinity-large-preview-free","n":"Trinity Large Preview","f":"trinity","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-06","rd":"2026-01-28","lu":"2026-01-28"},{"p":"opencode","id":"big-pickle","n":"Big Pickle","f":"big-pickle","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-10-17","lu":"2025-10-17"},{"p":"opencode","id":"gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"cr":30,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-24","lu":"2026-04-24"},{"p":"opencode","id":"qwen3.6-plus-free","n":"Qwen3.6 Plus Free","f":"qwen-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"opencode","id":"glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-11","lu":"2026-02-11"},{"p":"opencode","id":"mimo-v2-omni-free","n":"MiMo V2 Omni Free","f":"mimo-omni-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":262144,"out":64000,"mi":"tiap","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"opencode","id":"gpt-5-codex","n":"GPT-5 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.07,"co":8.5,"cr":0.107,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"opencode","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-14","lu":"2026-01-14"},{"p":"opencode","id":"glm-5-free","n":"GLM-5 Free","f":"glm-free","d":"Legacy model retained for compatibility with older integrations","ci":0,"co":0,"cr":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-11","lu":"2026-02-11"},{"p":"opencode","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen3.6","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"opencode","id":"nemotron-3-super-free","n":"Nemotron 3 Super Free","f":"nemotron-free","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0,"co":0,"cr":0,"ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-02","rd":"2026-03-11","lu":"2026-03-11"},{"p":"opencode","id":"grok-build-0.1","n":"Grok Build 0.1","f":"grok-build","d":"Grok coding model for agentic engineering, edits, and codebase workflows","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-20","lu":"2026-05-20"},{"p":"opencode","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.07,"co":8.5,"cr":0.107,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"opencode","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"inference","id":"mistral/mistral-nemo-12b-instruct","n":"Mistral Nemo 12B Instruct","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.038,"co":0.1,"ctx":16000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"google/gemma-3","n":"Google Gemma 3","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.15,"co":0.3,"ctx":125000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"osmosis/osmosis-structure-0.6b","n":"Osmosis Structure 0.6B","f":"osmosis","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.1,"co":0.5,"ctx":4000,"out":2048,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"qwen/qwen3-embedding-4b","n":"Qwen 3 Embedding 4B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.01,"co":0,"ctx":32000,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"qwen/qwen-2.5-7b-vision-instruct","n":"Qwen 2.5 7B Vision Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.2,"ctx":125000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"meta/llama-3.1-8b-instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.025,"co":0.025,"ctx":16000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"meta/llama-3.2-1b-instruct","n":"Llama 3.2 1B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.01,"co":0.01,"ctx":16000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"meta/llama-3.2-11b-vision-instruct","n":"Llama 3.2 11B Vision Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.055,"co":0.055,"ctx":16000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inference","id":"meta/llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.02,"co":0.02,"ctx":16000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"inceptron","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.66,"co":3.5,"cr":0.2,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"inceptron","id":"moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.75,"co":3.5,"cr":0.2,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"inceptron","id":"moonshotai/Kimi-K2.6-Fast","n":"Kimi K2.6 Fast","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":1.32,"co":7,"cr":0.4,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"inceptron","id":"zai-org/GLM-5.1-FP8","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":202752,"out":202752,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"inceptron","id":"zai-org/GLM-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.2,"co":4.2,"cr":0.26,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"inceptron","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.15,"co":0.9,"cr":0.05,"cw":0,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"llama","id":"llama-4-scout-17b-16e-instruct-fp8","n":"Llama-4-Scout-17B-16E-Instruct-FP8","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"llama","id":"cerebras-llama-4-maverick-17b-128e-instruct","n":"Cerebras-Llama-4-Maverick-17B-128E-Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-01","rd":"2025-04-05","lu":"2025-04-05"},{"p":"llama","id":"llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"llama","id":"groq-llama-4-maverick-17b-128e-instruct","n":"Groq-Llama-4-Maverick-17B-128E-Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-01","rd":"2025-04-05","lu":"2025-04-05"},{"p":"llama","id":"cerebras-llama-4-scout-17b-16e-instruct","n":"Cerebras-Llama-4-Scout-17B-16E-Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-01","rd":"2025-04-05","lu":"2025-04-05"},{"p":"llama","id":"llama-3.3-8b-instruct","n":"Llama-3.3-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"llama","id":"llama-4-maverick-17b-128e-instruct-fp8","n":"Llama-4-Maverick-17B-128E-Instruct-FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0,"co":0,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"llmtr","id":"sincap","n":"Sincap","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2026-05-05","lu":"2026-05-05"},{"p":"llmtr","id":"magibu-11b-v8","n":"Magibu 11B v8","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0,"co":0,"ctx":8192,"out":4096,"mi":"t","mo":"t","rd":"2026-06-05","lu":"2026-06-05"},{"p":"llmtr","id":"gemma-4","n":"Gemma 4","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":5,"co":10,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"llmtr","id":"medgemma-4b","n":"MedGemma 4B","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":3,"co":5,"ctx":8192,"out":4096,"mi":"ti","mo":"t","ow":1,"rd":"2026-04-26","lu":"2026-04-26"},{"p":"llmtr","id":"qwen3-6-35b","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":5,"co":10,"ctx":16384,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"llmtr","id":"trendyol-7b","n":"Trendyol 7B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2026-06-06","lu":"2026-06-06"},{"p":"cohere","id":"c4ai-aya-expanse-32b","n":"Aya Expanse 32B","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ctx":128000,"out":4000,"mi":"t","mo":"t","ow":1,"rd":"2024-10-24","lu":"2024-10-24"},{"p":"cohere","id":"command-a-03-2025","n":"Command A","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-03-13","lu":"2025-03-13"},{"p":"cohere","id":"c4ai-aya-vision-32b","n":"Aya Vision 32B","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ctx":16000,"out":4000,"mi":"ti","mo":"t","ow":1,"rd":"2025-03-04","lu":"2025-05-14"},{"p":"cohere","id":"command-r7b-arabic-02-2025","n":"Command R7B Arabic","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.0375,"co":0.15,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-02-27","lu":"2025-02-27"},{"p":"cohere","id":"c4ai-aya-vision-8b","n":"Aya Vision 8B","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ctx":16000,"out":4000,"mi":"ti","mo":"t","ow":1,"rd":"2025-03-04","lu":"2025-05-14"},{"p":"cohere","id":"command-r-08-2024","n":"Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.15,"co":0.6,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"cohere","id":"command-r7b-12-2024","n":"Command R7B","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.0375,"co":0.15,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-12-02","lu":"2024-12-02"},{"p":"cohere","id":"command-a-vision-07-2025","n":"Command A Vision","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":128000,"out":8000,"mi":"ti","mo":"t","ow":1,"kn":"2024-06-01","rd":"2025-07-31","lu":"2025-07-31"},{"p":"cohere","id":"command-a-plus-05-2026","n":"Command A Plus","f":"command-a","d":"Cohere's stronger command model for multilingual agents and enterprise workflows","ci":2.5,"co":10,"ctx":128000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04-01","rd":"2026-05-20","lu":"2026-06-09"},{"p":"cohere","id":"command-a-translate-08-2025","n":"Command A Translate","f":"command-a","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":2.5,"co":10,"ctx":8000,"out":8000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-08-28","lu":"2025-08-28"},{"p":"cohere","id":"command-r-plus-08-2024","n":"Command R+","f":"command-r","d":"Cohere's RAG workhorse for long-context enterprise search and tool use","ci":2.5,"co":10,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"cohere","id":"command-a-reasoning-08-2025","n":"Command A Reasoning","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-08-21","lu":"2025-08-21"},{"p":"cohere","id":"north-mini-code-1-0","n":"North Mini Code","f":"north","d":"Cohere coding model for practical software engineering and agentic edits","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09-23","rd":"2026-06-09","lu":"2026-06-09"},{"p":"cohere","id":"c4ai-aya-expanse-8b","n":"Aya Expanse 8B","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ctx":8000,"out":4000,"mi":"t","mo":"t","ow":1,"rd":"2024-10-24","lu":"2024-10-24"},{"p":"sarvam","id":"sarvam-105b","n":"Sarvam-105B","f":"sarvam","d":"Flagship Indian-language reasoning model for enterprise multilingual applications","ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-18","lu":"2026-03-06"},{"p":"sarvam","id":"sarvam-30b","n":"Sarvam-30B","f":"sarvam","d":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-18","lu":"2026-03-06"},{"p":"stepfun","id":"step-1-32k","n":"Step 1 (32K)","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":2.05,"co":9.59,"cr":0.41,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2025-01-01","lu":"2026-02-13"},{"p":"stepfun","id":"step-3.7-flash","n":"Step 3.7 Flash","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0.2,"co":1.15,"cr":0.04,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01-01","rd":"2026-05-29","lu":"2026-06-29"},{"p":"stepfun","id":"step-3.5-flash-2603","n":"Step 3.5 Flash 2603","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"cr":0.02,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-02","lu":"2026-04-02"},{"p":"stepfun","id":"stepaudio-2.5-tts","n":"StepAudio 2.5 TTS","f":"step","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2026-04-16","lu":"2026-07-02"},{"p":"stepfun","id":"stepaudio-2.5-asr","n":"StepAudio 2.5 ASR","f":"step","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ctx":0,"out":0,"mi":"a","mo":"t","rd":"2026-04-24","lu":"2026-07-02"},{"p":"stepfun","id":"step-3.5-flash","n":"Step 3.5 Flash","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.1,"co":0.3,"cr":0.02,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-06-15"},{"p":"stepfun","id":"step-tts-2","n":"Step TTS 2","f":"step","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","rd":"2026-03-01","lu":"2026-07-02"},{"p":"stepfun","id":"step-2-16k","n":"Step 2 (16K)","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":5.21,"co":16.44,"cr":1.04,"ctx":16384,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06","rd":"2025-01-01","lu":"2026-02-13"},{"p":"hpc-ai","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.3,"co":1.5,"cr":0.05,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01-01","rd":"2026-01-01","lu":"2026-06-01"},{"p":"hpc-ai","id":"zai-org/glm-5.1","n":"GLM 5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.615,"co":2.46,"cr":0.133,"ctx":202000,"out":202000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-08","lu":"2026-06-01"},{"p":"hpc-ai","id":"minimax/minimax-m2.5","n":"MiniMax M2.5","f":"minimax-m2.5","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-06-01"},{"p":"minimax-cn","id":"MiniMax-M2.1","n":"MiniMax-M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"minimax-cn","id":"MiniMax-M2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"minimax-cn","id":"MiniMax-M2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"minimax-cn","id":"MiniMax-M2","n":"MiniMax-M2","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":196608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"minimax-cn","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"minimax-cn","id":"MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":1000000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-25"},{"p":"minimax-cn","id":"MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"alibaba-coding-plan","id":"qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"alibaba-coding-plan","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"alibaba-coding-plan","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"alibaba-coding-plan","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.5,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"alibaba-coding-plan","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1875,"co":1.125,"cw":0.234375,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"alibaba-coding-plan","id":"qwen3-max-2026-01-23","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2026-01-23","lu":"2026-01-23"},{"p":"alibaba-coding-plan","id":"qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"alibaba-coding-plan","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"alibaba-coding-plan","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":196608,"out":24576,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-coding-plan","id":"qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-03","lu":"2026-02-03"},{"p":"alibaba-coding-plan","id":"glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"alibaba-coding-plan","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"longcat","id":"LongCat-2.0","n":"LongCat-2.0","f":"longcat","d":"Meituan LongCat-2.0, a reasoning model with tool calling and a 1M-token context window","ci":0.75,"co":2.95,"cr":0.015,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-06-30","lu":"2026-06-30"},{"p":"poe","id":"trytako/tako","n":"Tako","f":"tako","d":"Tool-capable chat model for instruction following and agentic application workflows","ctx":2048,"out":0,"mi":"t","mo":"t","tc":1,"rd":"2024-08-15","lu":"2024-08-15"},{"p":"poe","id":"xai/grok-code-fast-1","n":"Grok Code Fast 1","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":1.5,"cr":0.02,"ctx":256000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-22","lu":"2025-08-22"},{"p":"poe","id":"xai/grok-4.1-fast-reasoning","n":"Grok-4.1-Fast-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-19","lu":"2025-11-19"},{"p":"poe","id":"xai/grok-3-mini","n":"Grok 3 Mini","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.3,"co":0.5,"cr":0.075,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-11","lu":"2025-04-11"},{"p":"poe","id":"xai/grok-4.1-fast-non-reasoning","n":"Grok-4.1-Fast-Non-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ctx":2000000,"out":30000,"mi":"ti","mo":"t","tc":1,"rd":"2025-11-19","lu":"2025-11-19"},{"p":"poe","id":"xai/grok-3","n":"Grok 3","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-04-11","lu":"2025-04-11"},{"p":"poe","id":"xai/grok-4-fast-reasoning","n":"Grok-4-Fast-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-09-16","lu":"2025-09-16"},{"p":"poe","id":"xai/grok-4","n":"Grok-4","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"ctx":256000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-07-10","lu":"2025-07-10"},{"p":"poe","id":"xai/grok-4-fast-non-reasoning","n":"Grok-4-Fast-Non-Reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":128000,"mi":"ti","mo":"t","tc":1,"rd":"2025-09-16","lu":"2025-09-16"},{"p":"poe","id":"xai/grok-4.20-multi-agent","n":"Grok-4.20-Multi-Agent","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"cr":0.2,"ctx":128000,"out":0,"mi":"ti","mo":"t","tc":1,"rd":"2026-03-13","lu":"2026-03-13"},{"p":"poe","id":"topazlabs-co/topazlabs","n":"TopazLabs","f":"topazlabs","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":204,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2024-12-03","lu":"2024-12-03"},{"p":"poe","id":"fireworks-ai/kimi-k2.5-fw","n":"Kimi-K2.5-FW","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0,"co":0,"ctx":262144,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"poe","id":"google/veo-3.1-fast","n":"Veo-3.1-Fast","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":480,"out":0,"mi":"ti","mo":"v","tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"poe","id":"google/imagen-3","n":"Imagen-3","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2024-10-15","lu":"2024-10-15"},{"p":"poe","id":"google/nano-banana-pro","n":"Nano-Banana-Pro","f":"nano-banana","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":12,"cr":0.2,"ctx":65536,"out":0,"mi":"ti","mo":"i","tc":1,"rd":"2025-11-19","lu":"2025-11-19"},{"p":"poe","id":"google/lyria","n":"Lyria","f":"lyria","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":0,"out":0,"mi":"t","mo":"a","tc":1,"rd":"2025-06-04","lu":"2025-06-04"},{"p":"poe","id":"google/gemini-3.1-flash-lite","n":"Gemini-3.1-Flash-Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-02-18","lu":"2026-02-18"},{"p":"poe","id":"google/nano-banana","n":"Nano-Banana","f":"nano-banana","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.21,"co":1.8,"cr":0.021,"ctx":65536,"out":0,"mi":"ti","mo":"ti","tc":1,"rd":"2025-08-21","lu":"2025-08-21"},{"p":"poe","id":"google/gemini-deep-research","n":"gemini-deep-research","d":"Legacy model retained for compatibility with older integrations","ci":1.6,"co":9.6,"ctx":1048576,"out":0,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2025-12-11","lu":"2025-12-11"},{"p":"poe","id":"google/veo-3","n":"Veo-3","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":480,"out":0,"mi":"t","mo":"v","tc":1,"rd":"2025-05-21","lu":"2025-05-21"},{"p":"poe","id":"google/gemini-3-flash","n":"Gemini-3-Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.4,"co":2.4,"cr":0.04,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2025-10-07","lu":"2025-10-07"},{"p":"poe","id":"google/imagen-3-fast","n":"Imagen-3-Fast","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2024-10-17","lu":"2024-10-17"},{"p":"poe","id":"google/gemini-2.5-pro","n":"Gemini-2.5-Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":0.87,"co":7,"cr":0.087,"ctx":1065535,"out":65535,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2025-02-05","lu":"2025-02-05"},{"p":"poe","id":"google/veo-3.1","n":"Veo-3.1","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":480,"out":0,"mi":"t","mo":"v","tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"poe","id":"google/gemini-2.5-flash","n":"Gemini-2.5-Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.21,"co":1.8,"cr":0.021,"ctx":1065535,"out":65535,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2025-04-26","lu":"2025-04-26"},{"p":"poe","id":"google/imagen-4-fast","n":"Imagen-4-Fast","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2025-06-25","lu":"2025-06-25"},{"p":"poe","id":"google/gemini-3.5-flash","n":"Gemini-3.5-Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5152,"co":9.0909,"cr":0.1515,"ctx":1048576,"out":65536,"mi":"tia","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"poe","id":"google/gemma-4-31b","n":"Gemma-4-31B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":262144,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"poe","id":"google/veo-3-fast","n":"Veo-3-Fast","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":480,"out":0,"mi":"t","mo":"v","tc":1,"rd":"2025-10-13","lu":"2025-10-13"},{"p":"poe","id":"google/gemini-3-pro","n":"Gemini-3-Pro","f":"gemini-pro","d":"Legacy model retained for compatibility with older integrations","ci":1.6,"co":9.6,"cr":0.16,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2025-10-22","lu":"2025-10-22"},{"p":"poe","id":"google/gemini-2.0-flash","n":"Gemini-2.0-Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.1,"co":0.42,"ctx":990000,"out":8192,"mi":"tiva","mo":"t","tc":1,"rd":"2024-12-11","lu":"2024-12-11"},{"p":"poe","id":"google/gemini-2.5-flash-lite","n":"Gemini-2.5-Flash-Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.07,"co":0.28,"ctx":1024000,"out":64000,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2025-06-19","lu":"2025-06-19"},{"p":"poe","id":"google/imagen-4","n":"Imagen-4","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"poe","id":"google/imagen-4-ultra","n":"Imagen-4-Ultra","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2025-05-24","lu":"2025-05-24"},{"p":"poe","id":"google/gemini-3.1-pro","n":"Gemini-3.1-Pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-02-19","lu":"2026-02-19"},{"p":"poe","id":"google/veo-2","n":"Veo-2","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":480,"out":0,"mi":"t","mo":"v","tc":1,"rd":"2024-12-02","lu":"2024-12-02"},{"p":"poe","id":"google/gemini-2.0-flash-lite","n":"Gemini-2.0-Flash-Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.052,"co":0.21,"ctx":990000,"out":8192,"mi":"tiva","mo":"t","tc":1,"rd":"2025-02-05","lu":"2025-02-05"},{"p":"poe","id":"elevenlabs/elevenlabs-music","n":"ElevenLabs-Music","f":"elevenlabs","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":2000,"out":0,"mi":"t","mo":"a","tc":1,"rd":"2025-08-29","lu":"2025-08-29"},{"p":"poe","id":"elevenlabs/elevenlabs-v2.5-turbo","n":"ElevenLabs-v2.5-Turbo","f":"elevenlabs","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":128000,"out":0,"mi":"t","mo":"a","tc":1,"rd":"2024-10-28","lu":"2024-10-28"},{"p":"poe","id":"elevenlabs/elevenlabs-v3","n":"ElevenLabs-v3","f":"elevenlabs","d":"Speech generation model for controllable voice, narration, and audio delivery","ctx":128000,"out":0,"mi":"t","mo":"a","tc":1,"rd":"2025-06-05","lu":"2025-06-05"},{"p":"poe","id":"openai/chatgpt-4o-latest","n":"ChatGPT-4o-Latest","f":"gpt","d":"Legacy model retained for compatibility with older integrations","ci":4.5,"co":14,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2024-08-14","lu":"2024-08-14"},{"p":"poe","id":"openai/gpt-3.5-turbo-instruct","n":"GPT-3.5-Turbo-Instruct","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.4,"co":1.8,"ctx":3500,"out":1024,"mi":"ti","mo":"t","tc":1,"rd":"2023-09-20","lu":"2023-09-20"},{"p":"poe","id":"openai/o3","n":"o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.8,"co":7.2,"cr":0.45,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"poe","id":"openai/gpt-5.2-instant","n":"GPT-5.2-Instant","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.6,"co":13,"cr":0.16,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-12-11","lu":"2025-12-11"},{"p":"poe","id":"openai/gpt-4o-search","n":"GPT-4o-Search","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.2,"co":9,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-03-11","lu":"2025-03-11"},{"p":"poe","id":"openai/gpt-5.2-pro","n":"GPT-5.2-Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":19,"co":150,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-11","lu":"2025-12-11"},{"p":"poe","id":"openai/gpt-4-classic-0314","n":"GPT-4-Classic-0314","f":"gpt","d":"Legacy model retained for compatibility with older integrations","ci":27,"co":54,"ctx":8192,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-08-26","lu":"2024-08-26"},{"p":"poe","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.1,"co":9,"cr":0.11,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"poe","id":"openai/gpt-5-chat","n":"GPT-5-Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.1,"co":9,"cr":0.11,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-08-07","lu":"2025-08-07"},{"p":"poe","id":"openai/gpt-3.5-turbo","n":"GPT-3.5-Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.45,"co":1.4,"ctx":16384,"out":2048,"mi":"ti","mo":"t","tc":1,"rd":"2023-09-13","lu":"2023-09-13"},{"p":"poe","id":"openai/gpt-5-pro","n":"GPT-5-Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":14,"co":110,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-06","lu":"2025-10-06"},{"p":"poe","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2024-05-13","lu":"2024-05-13"},{"p":"poe","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0.99,"co":4,"cr":0.25,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"poe","id":"openai/sora-2","n":"Sora-2","f":"sora","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"ti","mo":"v","tc":1,"rd":"2025-10-06","lu":"2025-10-06"},{"p":"poe","id":"openai/o3-pro","n":"o3-pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":18,"co":72,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-06-10","lu":"2025-06-10"},{"p":"poe","id":"openai/gpt-4-classic","n":"GPT-4-Classic","f":"gpt","d":"Legacy model retained for compatibility with older integrations","ci":27,"co":54,"ctx":8192,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-03-25","lu":"2024-03-25"},{"p":"poe","id":"openai/gpt-4o-aug","n":"GPT-4o-Aug","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.2,"co":9,"cr":1.1,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2024-11-21","lu":"2024-11-21"},{"p":"poe","id":"openai/gpt-5.4-nano","n":"GPT-5.4-Nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.18,"co":1.1,"cr":0.018,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"poe","id":"openai/sora-2-pro","n":"Sora-2-Pro","f":"sora","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":0,"out":0,"mi":"ti","mo":"v","tc":1,"rd":"2025-10-06","lu":"2025-10-06"},{"p":"poe","id":"openai/gpt-5.1-codex","n":"GPT-5.1-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.1,"co":9,"cr":0.11,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-12","lu":"2025-11-12"},{"p":"poe","id":"openai/gpt-5.3-instant","n":"GPT-5.3-Instant","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.6,"co":13,"cr":0.16,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2026-03-03","lu":"2026-03-03"},{"p":"poe","id":"openai/gpt-5.3-codex-spark","n":"GPT-5.3-Codex-Spark","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0,"co":0,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-04","lu":"2026-03-04"},{"p":"poe","id":"openai/gpt-5.1-codex-max","n":"GPT-5.1-Codex-Max","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.1,"co":9,"cr":0.11,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-08","lu":"2025-12-08"},{"p":"poe","id":"openai/dall-e-3","n":"DALL-E-3","f":"dall-e","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":800,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2023-11-06","lu":"2023-11-06"},{"p":"poe","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0.99,"co":4,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-01-31","lu":"2025-01-31"},{"p":"poe","id":"openai/gpt-5.2","n":"GPT-5.2","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.6,"co":13,"cr":0.16,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-08","lu":"2025-12-08"},{"p":"poe","id":"openai/gpt-5.3-codex","n":"GPT-5.3-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.6,"co":13,"cr":0.16,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-10","lu":"2026-02-10"},{"p":"poe","id":"openai/gpt-5.1-codex-mini","n":"GPT-5.1-Codex-Mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.22,"co":1.8,"cr":0.022,"ctx":400000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-11-12","lu":"2025-11-12"},{"p":"poe","id":"openai/o4-mini-deep-research","n":"o4-mini-deep-research","f":"o-mini","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":1.8,"co":7.2,"cr":0.45,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-06-27","lu":"2025-06-27"},{"p":"poe","id":"openai/gpt-image-1.5","n":"gpt-image-1.5","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":128000,"out":0,"mi":"ti","mo":"i","rd":"2025-12-16","lu":"2025-12-16"},{"p":"poe","id":"openai/gpt-4.1-nano","n":"GPT-4.1-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.09,"co":0.36,"cr":0.022,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"poe","id":"openai/gpt-3.5-turbo-raw","n":"GPT-3.5-Turbo-Raw","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.45,"co":1.4,"ctx":4524,"out":2048,"mi":"ti","mo":"t","tc":1,"rd":"2023-09-27","lu":"2023-09-27"},{"p":"poe","id":"openai/o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":14,"co":54,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2024-12-18","lu":"2024-12-18"},{"p":"poe","id":"openai/o1-pro","n":"o1-pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":140,"co":540,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-03-19","lu":"2025-03-19"},{"p":"poe","id":"openai/gpt-5.4","n":"GPT-5.4","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2.2,"co":14,"cr":0.22,"ctx":1050000,"out":128000,"mi":"tip","mo":"i","rs":1,"tc":1,"rd":"2026-02-26","lu":"2026-02-26"},{"p":"poe","id":"openai/gpt-5.4-mini","n":"GPT-5.4-Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.68,"co":4,"cr":0.068,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-12","lu":"2026-03-12"},{"p":"poe","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.8,"co":7.2,"cr":0.45,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"rd":"2025-04-14","lu":"2025-04-14"},{"p":"poe","id":"openai/o3-deep-research","n":"o3-deep-research","f":"o","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":9,"co":36,"cr":2.2,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-06-27","lu":"2025-06-27"},{"p":"poe","id":"openai/gpt-5-mini","n":"GPT-5-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.22,"co":1.8,"cr":0.022,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-06-25","lu":"2025-06-25"},{"p":"poe","id":"openai/gpt-image-1","n":"GPT-Image-1","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":128000,"out":0,"mi":"ti","mo":"i","tc":1,"rd":"2025-03-31","lu":"2025-03-31"},{"p":"poe","id":"openai/gpt-4.1-mini","n":"GPT-4.1-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.36,"co":1.4,"cr":0.09,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"poe","id":"openai/gpt-4-turbo","n":"GPT-4-Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":9,"co":27,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2023-09-13","lu":"2023-09-13"},{"p":"poe","id":"openai/gpt-image-1-mini","n":"GPT-Image-1-Mini","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":0,"out":0,"mi":"ti","mo":"i","tc":1,"rd":"2025-08-26","lu":"2025-08-26"},{"p":"poe","id":"openai/gpt-5-nano","n":"GPT-5-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.045,"co":0.36,"cr":0.0045,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"poe","id":"openai/gpt-5.4-pro","n":"GPT-5.4-Pro","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":27,"co":160,"ctx":1050000,"out":128000,"mi":"ti","mo":"i","rs":1,"tc":1,"rd":"2026-03-05","lu":"2026-03-05"},{"p":"poe","id":"openai/o3-mini-high","n":"o3-mini-high","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0.99,"co":4,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-01-31","lu":"2025-01-31"},{"p":"poe","id":"openai/gpt-5.5-pro","n":"GPT-5.5-Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":27.2727,"co":163.6364,"ctx":400000,"out":128000,"mi":"ti","mo":"ti","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-08","lu":"2026-04-08"},{"p":"poe","id":"openai/gpt-4o-mini","n":"GPT-4o-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.14,"co":0.54,"cr":0.068,"ctx":124096,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-07-18","lu":"2024-07-18"},{"p":"poe","id":"openai/gpt-4o-mini-search","n":"GPT-4o-mini-Search","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.14,"co":0.54,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-03-11","lu":"2025-03-11"},{"p":"poe","id":"openai/gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.1,"co":9,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-09-23","lu":"2025-09-23"},{"p":"poe","id":"openai/gpt-5.2-codex","n":"GPT-5.2-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.6,"co":13,"cr":0.16,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01-14","lu":"2026-01-14"},{"p":"poe","id":"openai/gpt-5.1-instant","n":"GPT-5.1-Instant","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.1,"co":9,"cr":0.11,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-12","lu":"2025-11-12"},{"p":"poe","id":"openai/gpt-image-2","n":"GPT-Image-2","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":5.0505,"co":32.3232,"cr":1.2626,"ctx":0,"out":0,"mi":"ti","mo":"i","rd":"2026-04-21","lu":"2026-04-21"},{"p":"poe","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.1,"co":9,"cr":0.11,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-12","lu":"2025-11-12"},{"p":"poe","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":4.5455,"co":27.2727,"cr":0.4545,"ctx":400000,"out":128000,"mi":"ti","mo":"ti","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-08","lu":"2026-04-08"},{"p":"poe","id":"runwayml/runway","n":"Runway","f":"runway","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":256,"out":0,"mi":"ti","mo":"v","tc":1,"rd":"2024-10-11","lu":"2024-10-11"},{"p":"poe","id":"runwayml/runway-gen-4-turbo","n":"Runway-Gen-4-Turbo","f":"runway","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":256,"out":0,"mi":"ti","mo":"v","tc":1,"rd":"2025-05-09","lu":"2025-05-09"},{"p":"poe","id":"cerebras/llama-3.1-8b-cs","n":"Llama-3.1-8B-CS","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.1,"co":0.1,"ctx":128000,"out":0,"mi":"t","mo":"t","tc":1,"rd":"2025-05-13","lu":"2025-05-13"},{"p":"poe","id":"cerebras/qwen3-235b-2507-cs","n":"qwen3-235b-2507-cs","d":"Legacy model retained for compatibility with older integrations","ctx":0,"out":0,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"poe","id":"cerebras/gpt-oss-120b-cs","n":"GPT-OSS-120B-CS","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.35,"co":0.75,"ctx":128000,"out":0,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-06","lu":"2025-08-06"},{"p":"poe","id":"cerebras/llama-3.3-70b-cs","n":"llama-3.3-70b-cs","d":"Legacy model retained for compatibility with older integrations","ctx":0,"out":0,"mi":"t","mo":"t","rd":"2025-05-13","lu":"2025-05-13"},{"p":"poe","id":"cerebras/qwen3-32b-cs","n":"qwen3-32b-cs","d":"Legacy model retained for compatibility with older integrations","ctx":0,"out":0,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-15","lu":"2025-05-15"},{"p":"poe","id":"anthropic/claude-sonnet-3.5","n":"Claude-Sonnet-3.5","f":"claude-sonnet","d":"Legacy model retained for compatibility with older integrations","ci":2.6,"co":13,"cr":0.26,"cw":3.2,"ctx":189096,"out":8192,"mi":"tip","mo":"t","tc":1,"rd":"2024-06-05","lu":"2024-06-05"},{"p":"poe","id":"anthropic/claude-sonnet-3.7","n":"Claude-Sonnet-3.7","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.6,"co":13,"cr":0.26,"cw":3.2,"ctx":196608,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-02-19","lu":"2025-02-19"},{"p":"poe","id":"anthropic/claude-sonnet-3.5-june","n":"Claude-Sonnet-3.5-June","f":"claude-sonnet","d":"Legacy model retained for compatibility with older integrations","ci":2.6,"co":13,"cr":0.26,"cw":3.2,"ctx":189096,"out":8192,"mi":"tip","mo":"t","tc":1,"rd":"2024-11-18","lu":"2024-11-18"},{"p":"poe","id":"anthropic/claude-sonnet-4.5","n":"Claude-Sonnet-4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.6,"co":13,"cr":0.26,"cw":3.2,"ctx":983040,"out":32768,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-09-26","lu":"2025-09-26"},{"p":"poe","id":"anthropic/claude-sonnet-4","n":"Claude-Sonnet-4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.6,"co":13,"cr":0.26,"cw":3.2,"ctx":983040,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-21","lu":"2025-05-21"},{"p":"poe","id":"anthropic/claude-haiku-4.5","n":"Claude-Haiku-4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.85,"co":4.3,"cr":0.085,"cw":1.1,"ctx":192000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"poe","id":"anthropic/claude-haiku-3","n":"Claude-Haiku-3","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.21,"co":1.1,"cr":0.021,"cw":0.26,"ctx":189096,"out":8192,"mi":"tip","mo":"t","tc":1,"rd":"2024-03-09","lu":"2024-03-09"},{"p":"poe","id":"anthropic/claude-opus-4.7","n":"Claude-Opus-4.7","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.3,"co":21,"cr":0.43,"cw":5.4,"ctx":1048576,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-15","lu":"2026-04-15"},{"p":"poe","id":"anthropic/claude-haiku-3.5","n":"Claude-Haiku-3.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.68,"co":3.4,"cr":0.068,"cw":0.85,"ctx":189096,"out":8192,"mi":"tip","mo":"t","tc":1,"rd":"2024-10-01","lu":"2024-10-01"},{"p":"poe","id":"anthropic/claude-opus-4.8","n":"Claude-Opus-4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":4.2929,"co":21.4646,"ctx":1048576,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"poe","id":"anthropic/claude-opus-4.1","n":"Claude-Opus-4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":13,"co":64,"cr":1.3,"cw":16,"ctx":196608,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"poe","id":"anthropic/claude-opus-4.5","n":"Claude-Opus-4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.3,"co":21,"cr":0.43,"cw":5.3,"ctx":196608,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-11-21","lu":"2025-11-21"},{"p":"poe","id":"anthropic/claude-sonnet-4.6","n":"Claude-Sonnet-4.6","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.6,"co":13,"cr":0.26,"cw":3.2,"ctx":983040,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"poe","id":"anthropic/claude-opus-4","n":"Claude-Opus-4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":13,"co":64,"cr":1.3,"cw":16,"ctx":192512,"out":28672,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-21","lu":"2025-05-21"},{"p":"poe","id":"anthropic/claude-opus-4.6","n":"Claude-Opus-4.6","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.3,"co":21,"cr":0.43,"cw":5.3,"ctx":983040,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-04","lu":"2026-02-04"},{"p":"poe","id":"lumalabs/ray2","n":"Ray2","f":"ray","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":5000,"out":0,"mi":"ti","mo":"v","tc":1,"rd":"2025-02-20","lu":"2025-02-20"},{"p":"poe","id":"poetools/claude-code","n":"claude-code","d":"Claude model for careful reasoning, writing, coding, and tool use","ctx":0,"out":0,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-11-27","lu":"2025-11-27"},{"p":"poe","id":"empiriolabs/deepseek-v4-pro-el","n":"DeepSeek-V4-Pro-EL","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.67,"co":3.33,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-05-02"},{"p":"poe","id":"empiriolabs/deepseek-v4-flash-el","n":"DeepSeek-V4-Flash-EL","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-05-02"},{"p":"poe","id":"stabilityai/stablediffusionxl","n":"StableDiffusionXL","f":"stable-diffusion","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":200,"out":0,"mi":"ti","mo":"i","tc":1,"rd":"2023-07-09","lu":"2023-07-09"},{"p":"poe","id":"ideogramai/ideogram","n":"Ideogram","f":"ideogram","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":150,"out":0,"mi":"ti","mo":"i","tc":1,"rd":"2024-04-03","lu":"2024-04-03"},{"p":"poe","id":"ideogramai/ideogram-v2a","n":"Ideogram-v2a","f":"ideogram","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":150,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2025-02-27","lu":"2025-02-27"},{"p":"poe","id":"ideogramai/ideogram-v2","n":"Ideogram-v2","f":"ideogram","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":150,"out":0,"mi":"ti","mo":"i","tc":1,"rd":"2024-08-21","lu":"2024-08-21"},{"p":"poe","id":"ideogramai/ideogram-v2a-turbo","n":"Ideogram-v2a-Turbo","f":"ideogram","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":150,"out":0,"mi":"t","mo":"i","tc":1,"rd":"2025-02-27","lu":"2025-02-27"},{"p":"poe","id":"novita/glm-4.7","n":"glm-4.7","d":"Legacy model retained for compatibility with older integrations","ctx":205000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-22","lu":"2025-12-22"},{"p":"poe","id":"novita/minimax-m2.1","n":"minimax-m2.1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ctx":205000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-26","lu":"2025-12-26"},{"p":"poe","id":"novita/glm-4.6","n":"GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":0,"out":0,"mi":"t","mo":"t","tc":1,"rd":"2025-09-30","lu":"2025-09-30"},{"p":"poe","id":"novita/kimi-k2-thinking","n":"kimi-k2-thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ctx":256000,"out":0,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-11-07","lu":"2025-11-07"},{"p":"poe","id":"novita/kimi-k2.5","n":"Kimi-K2.5","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":128000,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"poe","id":"novita/glm-4.6v","n":"glm-4.6v","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ctx":131000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-09","lu":"2025-12-09"},{"p":"poe","id":"novita/kimi-k2.6","n":"Kimi-K2.6","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.96,"co":4.04,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-20","lu":"2026-05-02"},{"p":"poe","id":"novita/glm-4.7-flash","n":"glm-4.7-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ctx":200000,"out":65500,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"poe","id":"novita/glm-4.7-n","n":"glm-4.7-n","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ctx":205000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-22","lu":"2025-12-22"},{"p":"poe","id":"novita/glm-5","n":"GLM-5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":205000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"poe","id":"novita/deepseek-v3.2","n":"DeepSeek-V3.2","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.4,"cr":0.13,"ctx":128000,"out":0,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"kimi-for-coding","id":"k2p7","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"kimi-for-coding","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07","rd":"2025-11","lu":"2025-12"},{"p":"kimi-for-coding","id":"k2p5","n":"Kimi K2.5","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"kimi-for-coding","id":"k2p6","n":"Kimi K2.6","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04","lu":"2026-04"},{"p":"dinference","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.22,"co":0.88,"ctx":200000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"dinference","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.45,"co":1.65,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"dinference","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.25,"co":3.89,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"dinference","id":"gpt-oss-120b","n":"GPT OSS 120B","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.0675,"co":0.27,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-08","lu":"2025-08"},{"p":"dinference","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.75,"co":2.4,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"perplexity-agent","id":"xai/grok-4-1-fast-non-reasoning","n":"Grok 4.1 Fast (Non-Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","tc":1,"kn":"2025-07","rd":"2025-11-19","lu":"2025-11-19"},{"p":"perplexity-agent","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-05"},{"p":"perplexity-agent","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-05"},{"p":"perplexity-agent","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"perplexity-agent","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"perplexity-agent","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"perplexity-agent","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"perplexity-agent","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"perplexity-agent","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"perplexity-agent","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"perplexity-agent","id":"nvidia/nemotron-3-super-120b-a12b","n":"Nemotron 3 Super 120B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.25,"co":2.5,"ctx":1000000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-02","rd":"2026-03-11","lu":"2026-03-11"},{"p":"perplexity-agent","id":"anthropic/claude-opus-4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"perplexity-agent","id":"anthropic/claude-sonnet-4-5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"perplexity-agent","id":"anthropic/claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"perplexity-agent","id":"anthropic/claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"perplexity-agent","id":"anthropic/claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"perplexity-agent","id":"anthropic/claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"perplexity-agent","id":"perplexity/sonar","n":"Sonar","f":"sonar","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":0.25,"co":2.5,"cr":0.0625,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"siliconflow","id":"moonshotai/Kimi-K2.6","n":"moonshotai/Kimi-K2.6","f":"kimi","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.77,"co":4,"cr":0.2,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-21","lu":"2026-06-15"},{"p":"siliconflow","id":"moonshotai/Kimi-K2.5","n":"moonshotai/Kimi-K2.5","f":"kimi","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.45,"co":2.25,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"siliconflow","id":"baidu/ERNIE-4.5-300B-A47B","n":"baidu/ERNIE-4.5-300B-A47B","f":"ernie","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.28,"co":1.1,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-02","lu":"2025-11-25"},{"p":"siliconflow","id":"ByteDance-Seed/Seed-OSS-36B-Instruct","n":"ByteDance-Seed/Seed-OSS-36B-Instruct","f":"seed","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.21,"co":0.57,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-04","lu":"2025-11-25"},{"p":"siliconflow","id":"stepfun-ai/Step-3.5-Flash","n":"stepfun-ai/Step-3.5-Flash","f":"step","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"ctx":262000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"siliconflow","id":"google/gemma-4-31B-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.13,"co":0.4,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"siliconflow","id":"google/gemma-4-26B-A4B-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.12,"co":0.4,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"siliconflow","id":"inclusionAI/Ling-flash-2.0","n":"inclusionAI/Ling-flash-2.0","f":"ling","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.14,"co":0.57,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-18","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3.6-35B-A3B","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.2,"co":1.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"siliconflow","id":"Qwen/Qwen2.5-7B-Instruct","n":"Qwen/Qwen2.5-7B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.05,"co":0.05,"ctx":33000,"out":4000,"mi":"t","mo":"t","tc":1,"rd":"2024-09-18","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-235B-A22B-Instruct","n":"Qwen/Qwen3-VL-235B-A22B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-04","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3.6-27B","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":3.2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"siliconflow","id":"Qwen/Qwen3.5-397B-A17B","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.39,"co":2.34,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"siliconflow","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen/Qwen3-235B-A22B-Thinking-2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.13,"co":0.6,"ctx":262000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-07-28","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","n":"Qwen/Qwen3-Coder-480B-A35B-Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.25,"co":1,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-31","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3.5-122B-A10B","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.08,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"siliconflow","id":"Qwen/Qwen3-Coder-30B-A3B-Instruct","n":"Qwen/Qwen3-Coder-30B-A3B-Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.07,"co":0.28,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-01","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3.5-27B","n":"Qwen3.5 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.25,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"siliconflow","id":"Qwen/Qwen3-30B-A3B-Instruct-2507","n":"Qwen/Qwen3-30B-A3B-Instruct-2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.3,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-30","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-30B-A3B-Thinking","n":"Qwen/Qwen3-VL-30B-A3B-Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":1,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-11","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-8B-Instruct","n":"Qwen/Qwen3-VL-8B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.18,"co":0.68,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-15","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-235B-A22B-Thinking","n":"Qwen/Qwen3-VL-235B-A22B-Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.45,"co":3.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-04","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-8B","n":"Qwen/Qwen3-8B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.06,"co":0.06,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-30","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3.5-9B","n":"Qwen/Qwen3.5-9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-03-03","lu":"2026-04-24"},{"p":"siliconflow","id":"Qwen/Qwen3-32B","n":"Qwen/Qwen3-32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.14,"co":0.57,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-30","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-32B-Instruct","n":"Qwen/Qwen3-VL-32B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.6,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-21","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-14B","n":"Qwen/Qwen3-14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.07,"co":0.28,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-30","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen2.5-72B-Instruct","n":"Qwen/Qwen2.5-72B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.59,"co":0.59,"ctx":33000,"out":4000,"mi":"t","mo":"t","tc":1,"rd":"2024-09-18","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-30B-A3B-Instruct","n":"Qwen/Qwen3-VL-30B-A3B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.29,"co":1,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-05","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3-VL-32B-Thinking","n":"Qwen/Qwen3-VL-32B-Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-10-21","lu":"2025-11-25"},{"p":"siliconflow","id":"Qwen/Qwen3.5-35B-A3B","n":"Qwen3.5 35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.24,"co":1.8,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"siliconflow","id":"openai/gpt-oss-120b","n":"openai/gpt-oss-120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.45,"ctx":131000,"out":8000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-13","lu":"2025-11-25"},{"p":"siliconflow","id":"openai/gpt-oss-20b","n":"openai/gpt-oss-20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.04,"co":0.18,"ctx":131000,"out":8000,"mi":"t","mo":"t","tc":1,"rd":"2025-08-13","lu":"2025-11-25"},{"p":"siliconflow","id":"tencent/Hy3-preview","n":"Hy3 preview","f":"Hy","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.066,"co":0.26,"cr":0.029,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"siliconflow","id":"tencent/Hunyuan-A13B-Instruct","n":"tencent/Hunyuan-A13B-Instruct","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.14,"co":0.57,"ctx":131000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-06-30","lu":"2025-11-25"},{"p":"siliconflow","id":"zai-org/GLM-5V-Turbo","n":"zai-org/GLM-5V-Turbo","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":1.2,"co":4,"cw":0,"ctx":200000,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"siliconflow","id":"zai-org/GLM-5","n":"zai-org/GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.95,"co":2.55,"ctx":205000,"out":205000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-06-15"},{"p":"siliconflow","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cw":0,"ctx":1049000,"out":262000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"siliconflow","id":"zai-org/GLM-4.5-Air","n":"zai-org/GLM-4.5-Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.14,"co":0.86,"ctx":131000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2025-07-28","lu":"2025-11-25"},{"p":"siliconflow","id":"zai-org/GLM-5.1","n":"zai-org/GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cw":0,"ctx":205000,"out":205000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-08","lu":"2026-04-08"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-R1","n":"deepseek-ai/DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.18,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-05-28","lu":"2025-11-25"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V3.1-Terminus","n":"deepseek-ai/DeepSeek-V3.1-Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2025-11-25"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V3.1","n":"deepseek-ai/DeepSeek-V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-25","lu":"2025-11-25"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V3.2-Exp","n":"deepseek-ai/DeepSeek-V3.2-Exp","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.41,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-10-10","lu":"2025-11-25"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V4-Flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V4-Pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.74,"co":3.48,"cr":0.145,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V3.2","n":"deepseek-ai/DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.42,"ctx":164000,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-03","lu":"2025-12-03"},{"p":"siliconflow","id":"deepseek-ai/DeepSeek-V3","n":"deepseek-ai/DeepSeek-V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.25,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","tc":1,"rd":"2024-12-26","lu":"2025-11-25"},{"p":"siliconflow","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMaxAI/MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":197000,"out":131000,"mi":"t","mo":"t","tc":1,"rd":"2026-02-15","lu":"2026-06-15"},{"p":"umans-ai-coding-plan","id":"umans-kimi-k2.7","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"umans-ai-coding-plan","id":"umans-glm-5.1","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"umans-ai-coding-plan","id":"umans-coder","n":"Umans Coder","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"umans-ai-coding-plan","id":"umans-flash","n":"Umans Flash","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"umans-ai-coding-plan","id":"umans-glm-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"cr":0,"cw":0,"ctx":405504,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"umans-ai-coding-plan","id":"umans-qwen3.6-35b-a3b","n":"Qwen3.6 35B A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"io-net","id":"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8","n":"Llama 4 Maverick 17B 128E Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.15,"co":0.6,"cr":0.075,"cw":0.3,"ctx":430000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-15","lu":"2025-01-15"},{"p":"io-net","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.13,"co":0.38,"cr":0.065,"cw":0.26,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"io-net","id":"meta-llama/Llama-3.2-90B-Vision-Instruct","n":"Llama 3.2 90B Vision Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.35,"co":0.4,"cr":0.175,"cw":0.7,"ctx":16000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"io-net","id":"moonshotai/Kimi-K2-Thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.55,"co":2.25,"cr":0.275,"cw":1.1,"ctx":32768,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-08","rd":"2024-11-01","lu":"2024-11-01"},{"p":"io-net","id":"moonshotai/Kimi-K2-Instruct-0905","n":"Kimi K2 Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.39,"co":1.9,"cr":0.195,"cw":0.78,"ctx":32768,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-08","rd":"2024-09-05","lu":"2024-09-05"},{"p":"io-net","id":"Qwen/Qwen2.5-VL-32B-Instruct","n":"Qwen 2.5 VL 32B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.05,"co":0.22,"cr":0.025,"cw":0.1,"ctx":32000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-09","rd":"2024-11-01","lu":"2024-11-01"},{"p":"io-net","id":"Qwen/Qwen3-Next-80B-A3B-Instruct","n":"Qwen 3 Next 80B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.8,"cr":0.05,"cw":0.2,"ctx":262144,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-10","lu":"2025-01-10"},{"p":"io-net","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen 3 235B Thinking","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.11,"co":0.6,"cr":0.055,"cw":0.22,"ctx":262144,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-07-01","lu":"2025-07-01"},{"p":"io-net","id":"openai/gpt-oss-120b","n":"GPT-OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.04,"co":0.4,"cr":0.02,"cw":0.08,"ctx":131072,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-12-01","lu":"2024-12-01"},{"p":"io-net","id":"openai/gpt-oss-20b","n":"GPT-OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.03,"co":0.14,"cr":0.015,"cw":0.06,"ctx":64000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-12-01","lu":"2024-12-01"},{"p":"io-net","id":"mistralai/Devstral-Small-2505","n":"Devstral Small 2505","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.05,"co":0.22,"cr":0.025,"cw":0.1,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2025-05-01","lu":"2025-05-01"},{"p":"io-net","id":"mistralai/Magistral-Small-2506","n":"Magistral Small 2506","f":"magistral-small","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":0.5,"co":1.5,"cr":0.25,"cw":1,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2025-01","rd":"2025-06-01","lu":"2025-06-01"},{"p":"io-net","id":"mistralai/Mistral-Large-Instruct-2411","n":"Mistral Large Instruct 2411","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"cr":1,"cw":4,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2024-10","rd":"2024-11-01","lu":"2024-11-01"},{"p":"io-net","id":"mistralai/Mistral-Nemo-Instruct-2407","n":"Mistral Nemo Instruct 2407","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.02,"co":0.04,"cr":0.01,"cw":0.04,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-05","rd":"2024-07-01","lu":"2024-07-01"},{"p":"io-net","id":"Intel/Qwen3-Coder-480B-A35B-Instruct-int4-mixed-ar","n":"Qwen 3 Coder 480B","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.22,"co":0.95,"cr":0.11,"cw":0.44,"ctx":106000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-15","lu":"2025-01-15"},{"p":"io-net","id":"zai-org/GLM-4.6","n":"GLM 4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.4,"co":1.75,"cr":0.2,"cw":0.8,"ctx":200000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2024-11-15","lu":"2024-11-15"},{"p":"io-net","id":"deepseek-ai/DeepSeek-R1-0528","n":"DeepSeek R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":2,"co":8.75,"cr":1,"cw":4,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-28"},{"p":"gmicloud","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.855,"co":3.6,"cr":0.144,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"gmicloud","id":"moonshotai/kimi-k2.7-code-highspeed","n":"Kimi K2.7 Code Highspeed","f":"kimi-k2","d":"Lower-latency Kimi Code variant for interactive edits and coding-agent loops","ci":1.9,"co":8,"cr":0.38,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"gmicloud","id":"Qwen/Qwen3.7-Max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.25,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"gmicloud","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"gmicloud","id":"anthropic/claude-opus-4.7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":4.5,"co":22.5,"cr":0.45,"ctx":409600,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"gmicloud","id":"anthropic/claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"gmicloud","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"ctx":409600,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"gmicloud","id":"anthropic/claude-opus-4.6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"ctx":409600,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"gmicloud","id":"zai-org/GLM-5.1-FP8","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.98,"co":3.08,"cr":0.182,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"gmicloud","id":"zai-org/GLM-5-FP8","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.6,"co":1.92,"cr":0.12,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"gmicloud","id":"zai-org/GLM-5.2-FP8","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0.979,"co":3.08,"cr":0.182,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"gmicloud","id":"deepseek-ai/DeepSeek-V4-Flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.112,"co":0.224,"cr":0.022,"ctx":1048575,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"gmicloud","id":"deepseek-ai/DeepSeek-V4-Pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.392,"co":2.784,"cr":0.116,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2.5-tts-voiceclone","n":"MiMo-V2.5-TTS-VoiceClone","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2.5-tts-voicedesign","n":"MiMo-V2.5-TTS-VoiceDesign","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2-omni","n":"MiMo-V2-Omni","f":"mimo","d":"MiMo omni model for text, image, video, audio, and agents","ci":0,"co":0,"cr":0,"ctx":262144,"out":131072,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2-tts","n":"MiMo-V2-TTS","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2-pro","n":"MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-cn","id":"mimo-v2.5-tts","n":"MiMo-V2.5-TTS","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"zeldoc","id":"z-code","n":"Z-Code","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0,"co":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-04-15","lu":"2026-04-15"},{"p":"scaleway","id":"qwen3-235b-a22b-instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","ci":0.75,"co":2.25,"ctx":260000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-01","lu":"2026-03-17"},{"p":"scaleway","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3-Coder 30B-A3B Instruct","f":"qwen","d":"Smaller Qwen coder for efficient local agents and repo-level fixes","ci":0.2,"co":0.8,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2026-03-17"},{"p":"scaleway","id":"qwen3-embedding-8b","n":"Qwen3 Embedding 8B","f":"qwen","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":32768,"out":4096,"mi":"t","mo":"t","rd":"2025-25-11","lu":"2026-03-17"},{"p":"scaleway","id":"bge-multilingual-gemma2","n":"BGE Multilingual Gemma2","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.1,"co":0,"ctx":8191,"out":3072,"mi":"t","mo":"t","rd":"2024-07-26","lu":"2025-06-15"},{"p":"scaleway","id":"qwen3.6-35b-a3b","n":"Qwen3.6 35B A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.25,"co":1.5,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-05-01","lu":"2026-05-22"},{"p":"scaleway","id":"llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.9,"co":0.9,"ctx":100000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2026-03-17"},{"p":"scaleway","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.8,"co":5.5,"ctx":256000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"scaleway","id":"pixtral-12b-2409","n":"Pixtral 12B 2409","f":"pixtral","d":"Mistral vision-language model for image understanding and multimodal chat","ci":0.2,"co":0.2,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-09","rd":"2024-09-25","lu":"2026-03-17"},{"p":"scaleway","id":"mistral-small-3.2-24b-instruct-2506","n":"Mistral Small 3.2 24B Instruct (2506)","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.15,"co":0.35,"ctx":128000,"out":32768,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-03","rd":"2025-06-20","lu":"2026-03-17"},{"p":"scaleway","id":"gpt-oss-120b","n":"GPT-OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2024-01-01","lu":"2026-03-17"},{"p":"scaleway","id":"gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.25,"co":0.5,"ctx":256000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-01","lu":"2026-05-22"},{"p":"scaleway","id":"mistral-medium-3.5-128b","n":"Mistral Medium 3.5 128B","f":"mistral-medium","d":"Balanced Mistral model for enterprise assistants, multilingual work, and tools","ci":1.5,"co":7.5,"ctx":256000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"scaleway","id":"qwen3.5-397b-a17b","n":"Qwen3.5 397B A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.6,"co":3.6,"ctx":256000,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-03-17","lu":"2026-03-17"},{"p":"scaleway","id":"whisper-large-v3","n":"Whisper Large v3","f":"whisper","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0.003,"co":0,"ctx":0,"out":8192,"mi":"a","mo":"t","ow":1,"kn":"2023-09","rd":"2023-09-01","lu":"2026-03-17"},{"p":"scaleway","id":"gemma-3-27b-it","n":"Gemma-3-27B-IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.25,"co":0.5,"ctx":40000,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2024-12-01","lu":"2026-03-17"},{"p":"scaleway","id":"voxtral-small-24b-2507","n":"Voxtral Small 24B 2507","f":"voxtral","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.15,"co":0.35,"ctx":32000,"out":16384,"mi":"ta","mo":"t","tc":1,"ow":1,"rd":"2025-07-01","lu":"2026-03-17"},{"p":"scaleway","id":"devstral-2-123b-instruct-2512","n":"Devstral 2 123B Instruct (2512)","f":"devstral","d":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","ci":0.4,"co":2,"ctx":256000,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2026-01-07","lu":"2026-03-17"},{"p":"ovhcloud","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3-Coder-30B-A3B-Instruct","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.07,"co":0.26,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-10-28","lu":"2025-10-28"},{"p":"ovhcloud","id":"qwen3-32b","n":"Qwen3-32B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.09,"co":0.25,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-16","lu":"2025-07-16"},{"p":"ovhcloud","id":"qwen3guard-gen-8b","n":"Qwen3Guard-Gen-8B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ctx":32768,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2026-01-22","lu":"2026-01-22"},{"p":"ovhcloud","id":"qwen3guard-gen-0.6b","n":"Qwen3Guard-Gen-0.6B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ctx":32768,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2026-01-22","lu":"2026-01-22"},{"p":"ovhcloud","id":"meta-llama-3_3-70b-instruct","n":"Meta-Llama-3_3-70B-Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.74,"co":0.74,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-01","lu":"2025-04-01"},{"p":"ovhcloud","id":"mistral-small-3.2-24b-instruct-2506","n":"Mistral-Small-3.2-24B-Instruct-2506","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.31,"ctx":131072,"out":131072,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-07-16","lu":"2025-07-16"},{"p":"ovhcloud","id":"qwen2.5-vl-72b-instruct","n":"Qwen2.5-VL-72B-Instruct","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":1.01,"co":1.01,"ctx":32768,"out":32768,"mi":"ti","mo":"t","ow":1,"rd":"2025-03-31","lu":"2025-03-31"},{"p":"ovhcloud","id":"gpt-oss-120b","n":"gpt-oss-120b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.09,"co":0.47,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-28","lu":"2025-08-28"},{"p":"ovhcloud","id":"mistral-7b-instruct-v0.3","n":"Mistral-7B-Instruct-v0.3","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.11,"co":0.11,"ctx":65536,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-01","lu":"2025-04-01"},{"p":"ovhcloud","id":"mistral-nemo-instruct-2407","n":"Mistral-Nemo-Instruct-2407","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.14,"co":0.14,"ctx":65536,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-20","lu":"2024-11-20"},{"p":"ovhcloud","id":"qwen3.6-27b","n":"Qwen3.6-27B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.47,"co":3.19,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"ovhcloud","id":"qwen3.5-9b","n":"Qwen3.5-9B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.12,"co":0.18,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"ovhcloud","id":"qwen3.5-397b-a17b","n":"Qwen3.5-397B-A17B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.71,"co":4.25,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-05-18","lu":"2026-05-18"},{"p":"ovhcloud","id":"gpt-oss-20b","n":"gpt-oss-20b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.18,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-28","lu":"2025-08-28"},{"p":"friendli","id":"google/gemma-4-31B-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.14,"co":0.4,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"friendli","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","ci":0.2,"co":0.8,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-29","lu":"2026-01-29"},{"p":"friendli","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"friendli","id":"zai-org/GLM-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"ctx":202752,"out":202752,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"friendli","id":"deepseek-ai/DeepSeek-V3.2","n":"DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.5,"co":1.5,"cr":0.25,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"friendli","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.3,"co":1.2,"cr":0.06,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"tencent-tokenhub","id":"hy3","n":"Hy3","f":"Hy","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-07-06","lu":"2026-07-06"},{"p":"tencent-tokenhub","id":"hy3-preview","n":"Hy3 preview","f":"Hy","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"wandb","id":"meta-llama/Llama-4-Scout-17B-16E-Instruct","n":"Llama 4 Scout 17B 16E Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.17,"co":0.66,"ctx":64000,"out":64000,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-31","lu":"2026-03-12"},{"p":"wandb","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.71,"co":0.71,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2026-03-12"},{"p":"wandb","id":"meta-llama/Llama-3.1-70B-Instruct","n":"Llama 3.1 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.8,"co":0.8,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-23","lu":"2026-03-12"},{"p":"wandb","id":"meta-llama/Llama-3.1-8B-Instruct","n":"Meta-Llama-3.1-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.22,"co":0.22,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2026-03-12"},{"p":"wandb","id":"moonshotai/Kimi-K2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.5,"co":2.85,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-03-12"},{"p":"wandb","id":"microsoft/Phi-4-mini-instruct","n":"Phi-4-mini-instruct","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.08,"co":0.35,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2026-03-12"},{"p":"wandb","id":"Qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen3-235B-A22B-Thinking-2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.1,"co":0.1,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-25","lu":"2026-03-12"},{"p":"wandb","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct","n":"Qwen3-Coder-480B-A35B-Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":1,"co":1.5,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2026-03-12"},{"p":"wandb","id":"Qwen/Qwen3-30B-A3B-Instruct-2507","n":"Qwen3 30B A3B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.3,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-29","lu":"2026-03-12"},{"p":"wandb","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.1,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04-28","lu":"2026-03-12"},{"p":"wandb","id":"openai/gpt-oss-120b","n":"gpt-oss-120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2026-03-12"},{"p":"wandb","id":"openai/gpt-oss-20b","n":"gpt-oss-20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.2,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2026-03-12"},{"p":"wandb","id":"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8","n":"NVIDIA Nemotron 3 Super 120B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.2,"co":0.8,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-12"},{"p":"wandb","id":"OpenPipe/Qwen3-14B-Instruct","n":"OpenPipe Qwen3 14B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.05,"co":0.22,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04-29","lu":"2026-03-12"},{"p":"wandb","id":"zai-org/GLM-5-FP8","n":"GLM 5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"ctx":200000,"out":200000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-03-12"},{"p":"wandb","id":"zai-org/GLM-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.26,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"wandb","id":"deepseek-ai/DeepSeek-V3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.55,"co":1.65,"ctx":161000,"out":161000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-08-21","lu":"2026-03-12"},{"p":"wandb","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-03-12"},{"p":"kuae-cloud-coding-plan","id":"GLM-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"gitlab","id":"duo-chat-opus-4-5","n":"Agentic Chat (Claude Opus 4.5)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2026-01-08","lu":"2026-01-08"},{"p":"gitlab","id":"duo-chat-opus-4-8","n":"Agentic Chat (Claude Opus 4.8)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-05-28","lu":"2026-05-28"},{"p":"gitlab","id":"duo-chat-opus-4-7","n":"Agentic Chat (Claude Opus 4.7)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"gitlab","id":"duo-chat-gpt-5-2-codex","n":"Agentic Chat (GPT-5.2 Codex)","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-22","lu":"2026-01-22"},{"p":"gitlab","id":"duo-chat-fable-5","n":"Agentic Chat (Claude Fable 5)","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"gitlab","id":"duo-chat-gpt-5-5","n":"Agentic Chat (GPT-5.5)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-04-23","lu":"2026-04-23"},{"p":"gitlab","id":"duo-chat-opus-4-6","n":"Agentic Chat (Claude Opus 4.6)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"gitlab","id":"duo-chat-gpt-5-4","n":"Agentic Chat (GPT-5.4)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"gitlab","id":"duo-chat-gpt-5-codex","n":"Agentic Chat (GPT-5 Codex)","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2026-01-22","lu":"2026-01-22"},{"p":"gitlab","id":"duo-chat-gpt-5-4-nano","n":"Agentic Chat (GPT-5.4 Nano)","f":"gpt-nano","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"gitlab","id":"duo-chat-sonnet-4-6","n":"Agentic Chat (Claude Sonnet 4.6)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"gitlab","id":"duo-chat-gpt-5-mini","n":"Agentic Chat (GPT-5 Mini)","f":"gpt-mini","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2026-01-22","lu":"2026-01-22"},{"p":"gitlab","id":"duo-chat-sonnet-5","n":"Agentic Chat (Claude Sonnet 5)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"gitlab","id":"duo-chat-gpt-5-4-mini","n":"Agentic Chat (GPT-5.4 Mini)","f":"gpt-mini","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"gitlab","id":"duo-chat-gpt-5-3-codex","n":"Agentic Chat (GPT-5.3 Codex)","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"gitlab","id":"duo-chat-haiku-4-5","n":"Agentic Chat (Claude Haiku 4.5)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2026-01-08","lu":"2026-01-08"},{"p":"gitlab","id":"duo-chat-gpt-5-2","n":"Agentic Chat (GPT-5.2)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-23","lu":"2026-01-23"},{"p":"gitlab","id":"duo-chat-sonnet-4-5","n":"Agentic Chat (Claude Sonnet 4.5)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2026-01-08","lu":"2026-01-08"},{"p":"gitlab","id":"duo-chat-gpt-5-1","n":"Agentic Chat (GPT-5.1)","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0,"co":0,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2026-01-22","lu":"2026-01-22"},{"p":"kilo","id":"inclusionai/ling-2.6-1t","n":"inclusionAI: Ling-2.6-1T","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.3,"co":2.5,"cr":0.06,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-04-23","lu":"2026-05-16"},{"p":"kilo","id":"inclusionai/ring-2.6-1t","n":"inclusionAI: Ring-2.6-1T","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.075,"co":0.625,"cr":0.015,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-08","lu":"2026-05-16"},{"p":"kilo","id":"inclusionai/ling-2.6-flash","n":"inclusionAI: Ling-2.6 Flash","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.08,"co":0.24,"cr":0.016,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-04-21","lu":"2026-05-01"},{"p":"kilo","id":"ibm-granite/granite-4.0-h-micro","n":"IBM: Granite 4.0 Micro","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.017,"co":0.11,"ctx":131000,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2025-10-20","lu":"2026-03-15"},{"p":"kilo","id":"ibm-granite/granite-4.1-8b","n":"IBM: Granite 4.1 8B","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.05,"co":0.1,"cr":0.05,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2026-04-30","lu":"2026-05-01"},{"p":"kilo","id":"meta-llama/llama-3.1-8b-instruct","n":"Meta: Llama 3.1 8B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.02,"co":0.05,"ctx":16384,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-23","lu":"2025-12-23"},{"p":"kilo","id":"meta-llama/llama-3-70b-instruct","n":"Meta: Llama 3 70B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.51,"co":0.74,"ctx":8192,"out":8000,"mi":"t","mo":"t","ow":1,"rd":"2024-07-23","lu":"2024-07-23"},{"p":"kilo","id":"meta-llama/llama-3.1-70b-instruct","n":"Meta: Llama 3.1 70B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.4,"co":0.4,"ctx":131072,"out":26215,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-16","lu":"2024-07-23"},{"p":"kilo","id":"meta-llama/llama-3.2-1b-instruct","n":"Meta: Llama 3.2 1B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.027,"co":0.2,"ctx":60000,"out":12000,"mi":"t","mo":"t","ow":1,"rd":"2024-09-18","lu":"2026-01-27"},{"p":"kilo","id":"meta-llama/llama-4-maverick","n":"Meta: Llama 4 Maverick","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.15,"co":0.6,"ctx":1048576,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-04-05","lu":"2025-12-24"},{"p":"kilo","id":"meta-llama/llama-3.2-11b-vision-instruct","n":"Meta: Llama 3.2 11B Vision Instruct","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.049,"co":0.049,"ctx":131072,"out":16384,"mi":"ti","mo":"t","ow":1,"rd":"2024-09-25","lu":"2024-09-25"},{"p":"kilo","id":"meta-llama/llama-3.3-70b-instruct","n":"Meta: Llama 3.3 70B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.1,"co":0.32,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-08-01","lu":"2026-02-04"},{"p":"kilo","id":"meta-llama/llama-guard-3-8b","n":"Llama Guard 3 8B","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.02,"co":0.06,"ctx":131072,"out":26215,"mi":"t","mo":"t","ow":1,"rd":"2024-04-18","lu":"2026-02-04"},{"p":"kilo","id":"meta-llama/llama-guard-4-12b","n":"Meta: Llama Guard 4 12B","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.18,"co":0.18,"ctx":163840,"out":32768,"mi":"it","mo":"t","ow":1,"rd":"2025-04-05","lu":"2025-04-05"},{"p":"kilo","id":"meta-llama/llama-3-8b-instruct","n":"Meta: Llama 3 8B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.03,"co":0.04,"ctx":8192,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-04-25","lu":"2025-04-03"},{"p":"kilo","id":"meta-llama/llama-4-scout","n":"Meta: Llama 4 Scout","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.08,"co":0.3,"ctx":327680,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-04-05","lu":"2025-04-05"},{"p":"kilo","id":"meta-llama/llama-3.2-3b-instruct","n":"Meta: Llama 3.2 3B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.051,"co":0.34,"ctx":80000,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2024-09-18","lu":"2026-03-15"},{"p":"kilo","id":"~anthropic/claude-haiku-latest","n":"Anthropic: Claude Haiku Latest","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"~anthropic/claude-sonnet-latest","n":"Anthropic: Claude Sonnet Latest","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"~anthropic/claude-opus-latest","n":"Anthropic: Claude Opus Latest","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-05-01"},{"p":"kilo","id":"kilo-auto/balanced","n":"Kilo Auto Balanced","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.6,"co":3,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"kilo","id":"kilo-auto/small","n":"Kilo Auto Small","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.05,"co":0.4,"ctx":400000,"out":128000,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"kilo","id":"kilo-auto/free","n":"Kilo Auto Free","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0,"co":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"kilo","id":"kilo-auto/frontier","n":"Kilo Auto Frontier","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":5,"co":25,"ctx":1000000,"out":128000,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"kilo","id":"moonshotai/kimi-k2","n":"MoonshotAI: Kimi K2 0711","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.55,"co":2.2,"ctx":131000,"out":26215,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-11","lu":"2026-03-15"},{"p":"kilo","id":"moonshotai/kimi-k2-thinking","n":"MoonshotAI: Kimi K2 Thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.47,"co":2,"cr":0.2,"ctx":131072,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-06","lu":"2026-03-15"},{"p":"kilo","id":"moonshotai/kimi-k2.5","n":"MoonshotAI: Kimi K2.5","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.45,"co":2.2,"ctx":262144,"out":65535,"mi":"it","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-27","lu":"2026-03-15"},{"p":"kilo","id":"moonshotai/kimi-k2.6","n":"MoonshotAI: Kimi K2.6","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.75,"co":3.5,"cr":0.375,"ctx":262144,"out":65535,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-20","lu":"2026-05-12"},{"p":"kilo","id":"moonshotai/kimi-k2-0905","n":"MoonshotAI: Kimi K2 0905","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.4,"co":2,"cr":0.15,"ctx":131072,"out":26215,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-09-05","lu":"2025-09-05"},{"p":"kilo","id":"baidu/ernie-4.5-300b-a47b","n":"Baidu: ERNIE 4.5 300B A47B ","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.28,"co":1.1,"ctx":123000,"out":12000,"mi":"t","mo":"t","ow":1,"rd":"2025-06-30","lu":"2026-01"},{"p":"kilo","id":"baidu/ernie-4.5-vl-28b-a3b","n":"Baidu: ERNIE 4.5 VL 28B A3B","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.14,"co":0.56,"ctx":30000,"out":8000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-30","lu":"2025-06-30"},{"p":"kilo","id":"baidu/ernie-4.5-vl-424b-a47b","n":"Baidu: ERNIE 4.5 VL 424B A47B ","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.42,"co":1.25,"ctx":123000,"out":16000,"mi":"it","mo":"t","rs":1,"ow":1,"rd":"2025-06-30","lu":"2026-01"},{"p":"kilo","id":"baidu/ernie-4.5-21b-a3b-thinking","n":"Baidu: ERNIE 4.5 21B A3B Thinking","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.07,"co":0.28,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-09-19","lu":"2025-09-19"},{"p":"kilo","id":"baidu/cobuddy:free","n":"Baidu: CoBuddy (free)","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-06","lu":"2026-05-07"},{"p":"kilo","id":"baidu/ernie-4.5-21b-a3b","n":"Baidu: ERNIE 4.5 21B A3B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.07,"co":0.28,"ctx":120000,"out":8000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-06-30","lu":"2025-06-30"},{"p":"kilo","id":"baidu/qianfan-ocr-fast","n":"Baidu: Qianfan-OCR-Fast","d":"OCR model for extracting structured text from documents and screenshots","ci":0.68,"co":2.81,"ctx":65536,"out":28672,"mi":"it","mo":"t","rd":"2026-04-20","lu":"2026-05-16"},{"p":"kilo","id":"perceptron/perceptron-mk1","n":"Perceptron: Perceptron Mk1","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.15,"co":1.5,"ctx":32768,"out":8192,"mi":"itv","mo":"t","rs":1,"rd":"2026-05-12","lu":"2026-05-16"},{"p":"kilo","id":"alfredpros/codellama-7b-instruct-solidity","n":"AlfredPros: CodeLLaMa 7B Instruct Solidity","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.8,"co":1.2,"ctx":4096,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2025-04-14","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-3.1-flash-lite","n":"Google: Gemini 3.1 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":0.08333,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2026-05-07","lu":"2026-05-16"},{"p":"kilo","id":"google/gemma-3n-e4b-it","n":"Google: Gemma 3n 4B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.02,"co":0.04,"ctx":32768,"out":6554,"mi":"t","mo":"t","ow":1,"rd":"2025-05-20","lu":"2025-05-20"},{"p":"kilo","id":"google/gemini-2.5-pro","n":"Google: Gemini 2.5 Pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"cw":0.375,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2025-03-20","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-2.5-flash","n":"Google: Gemini 2.5 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.03,"cw":0.083333,"ctx":1048576,"out":65535,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2025-07-17","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-3.5-flash","n":"Google: Gemini 3.5 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"cw":0.08333,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2026-05-19","lu":"2026-05-27"},{"p":"kilo","id":"google/gemini-2.0-flash-lite-001","n":"Google: Gemini 2.0 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.075,"co":0.3,"ctx":1048576,"out":8192,"mi":"aiptv","mo":"t","tc":1,"rd":"2024-12-11","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-2.5-flash-lite-preview-09-2025","n":"Google: Gemini 2.5 Flash Lite Preview 09-2025","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.01,"cw":0.083333,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2025-09-25","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-2.0-flash-001","n":"Google: Gemini 2.0 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.1,"co":0.4,"cr":0.025,"cw":0.083333,"ctx":1048576,"out":8192,"mi":"aiptv","mo":"t","tc":1,"rd":"2024-12-11","lu":"2026-03-15"},{"p":"kilo","id":"google/gemma-4-31b-it","n":"Google: Gemma 4 31B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.14,"co":0.4,"ctx":262144,"out":131072,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-11"},{"p":"kilo","id":"google/lyria-3-clip-preview","n":"Google: Lyria 3 Clip Preview","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":1048576,"out":65536,"mi":"it","mo":"at","rd":"2026-03-30","lu":"2026-04-11"},{"p":"kilo","id":"google/gemini-3.1-pro-preview-customtools","n":"Google: Gemini 3.1 Pro Preview Custom Tools","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-3-pro-image-preview","n":"Google: Nano Banana Pro (Gemini 3 Pro Image Preview)","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":12,"ctx":65536,"out":32768,"mi":"it","mo":"it","rs":1,"rd":"2025-11-20","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-2.5-flash-image","n":"Google: Nano Banana (Gemini 2.5 Flash Image)","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.3,"co":2.5,"ctx":32768,"out":32768,"mi":"it","mo":"it","rd":"2025-10-08","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-2.5-flash-lite","n":"Google: Gemini 2.5 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.01,"cw":0.083333,"ctx":1048576,"out":65535,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2025-06-17","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-3.1-flash-image-preview","n":"Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":3,"ctx":65536,"out":65536,"mi":"it","mo":"it","rs":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-2.5-pro-preview-05-06","n":"Google: Gemini 2.5 Pro Preview 05-06","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"cw":0.375,"ctx":1048576,"out":65535,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2025-05-06","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-3.1-pro-preview","n":"Google: Gemini 3.1 Pro Preview","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2026-02-19","lu":"2026-03-15"},{"p":"kilo","id":"google/gemma-4-26b-a4b-it","n":"Google: Gemma 4 26B A4B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.12,"co":0.4,"ctx":262144,"out":262144,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-03","lu":"2026-04-11"},{"p":"kilo","id":"google/gemini-2.5-pro-preview","n":"Google: Gemini 2.5 Pro Preview 06-05","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"cw":0.375,"ctx":1048576,"out":65536,"mi":"aipt","mo":"t","rs":1,"tc":1,"rd":"2025-06-05","lu":"2026-03-15"},{"p":"kilo","id":"google/gemini-3-flash-preview","n":"Google: Gemini 3 Flash Preview","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"cw":0.083333,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2025-12-17","lu":"2026-03-15"},{"p":"kilo","id":"google/gemma-3-12b-it","n":"Google: Gemma 3 12B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.04,"co":0.13,"cr":0.015,"ctx":131072,"out":131072,"mi":"it","mo":"t","ow":1,"rd":"2025-03-13","lu":"2026-03-15"},{"p":"kilo","id":"google/gemma-3-4b-it","n":"Google: Gemma 3 4B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.04,"co":0.08,"ctx":131072,"out":19200,"mi":"it","mo":"t","ow":1,"rd":"2025-03-13","lu":"2026-03-15"},{"p":"kilo","id":"google/gemma-3-27b-it","n":"Google: Gemma 3 27B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.03,"co":0.11,"cr":0.02,"ctx":128000,"out":65536,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2025-03-12","lu":"2026-03-15"},{"p":"kilo","id":"google/lyria-3-pro-preview","n":"Google: Lyria 3 Pro Preview","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":1048576,"out":65536,"mi":"it","mo":"at","rd":"2026-03-30","lu":"2026-04-11"},{"p":"kilo","id":"google/gemma-2-27b-it","n":"Google: Gemma 2 27B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.65,"co":0.65,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2024-06-24","lu":"2024-06-24"},{"p":"kilo","id":"google/gemini-3.1-flash-lite-preview","n":"Google: Gemini 3.1 Flash Lite Preview","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"ctx":1048576,"out":65536,"mi":"aiptv","mo":"t","rs":1,"tc":1,"rd":"2026-03-03","lu":"2026-03-15"},{"p":"kilo","id":"liquid/lfm-2-24b-a2b","n":"LiquidAI: LFM2-24B-A2B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.03,"co":0.12,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"x-ai/grok-4.20","n":"xAI: Grok 4.20","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-03-31","lu":"2026-04-11"},{"p":"kilo","id":"x-ai/grok-4.3","n":"xAI: Grok 4.3","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":4096,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"kilo","id":"x-ai/grok-4.20-multi-agent","n":"xAI: Grok 4.20 Multi-Agent","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"ipt","mo":"t","rs":1,"rd":"2026-03-31","lu":"2026-04-11"},{"p":"kilo","id":"x-ai/grok-build-0.1","n":"xAI: Grok Build 0.1","d":"Grok coding model for agentic engineering, edits, and codebase workflows","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2026-05-20","lu":"2026-05-27"},{"p":"kilo","id":"~google/gemini-pro-latest","n":"Google: Gemini Pro Latest","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"cw":0.375,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"~google/gemini-flash-latest","n":"Google: Gemini Flash Latest","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"cw":0.08333333333333334,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"microsoft/phi-4-mini-instruct","n":"Microsoft: Phi 4 Mini Instruct","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.08,"co":0.35,"cr":0.08,"ctx":128000,"out":128000,"mi":"t","mo":"t","ow":1,"rd":"2025-10-17","lu":"2026-05-07"},{"p":"kilo","id":"microsoft/phi-4","n":"Microsoft: Phi 4","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.06,"co":0.14,"ctx":16384,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2024-12-11","lu":"2024-12-11"},{"p":"kilo","id":"microsoft/wizardlm-2-8x22b","n":"WizardLM-2 8x22B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.62,"co":0.62,"ctx":65535,"out":8000,"mi":"t","mo":"t","ow":1,"rd":"2024-04-24","lu":"2024-04-24"},{"p":"kilo","id":"poolside/laguna-xs.2:free","n":"Poolside: Laguna XS.2 (free)","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-06-13"},{"p":"kilo","id":"poolside/laguna-m.1:free","n":"Poolside: Laguna M.1 (free)","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-06-13"},{"p":"kilo","id":"writer/palmyra-x5","n":"Writer: Palmyra X5","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.6,"co":6,"ctx":1040000,"out":8192,"mi":"t","mo":"t","rd":"2025-04-28","lu":"2025-04-28"},{"p":"kilo","id":"z-ai/glm-4.7","n":"Z.ai: GLM 4.7","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.38,"co":1.98,"cr":0.2,"ctx":202752,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-22","lu":"2026-03-15"},{"p":"kilo","id":"z-ai/glm-4.5v","n":"Z.ai: GLM 4.5V","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"cr":0.11,"ctx":65536,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-11","lu":"2025-08-11"},{"p":"kilo","id":"z-ai/glm-4.5","n":"Z.ai: GLM 4.5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.175,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-28","lu":"2026-03-15"},{"p":"kilo","id":"z-ai/glm-5.1","n":"Z.ai: GLM 5.1","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.26,"co":3.96,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"kilo","id":"z-ai/glm-4.6","n":"Z.ai: GLM 4.6","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.39,"co":1.9,"cr":0.175,"ctx":204800,"out":204800,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-30","lu":"2026-03-15"},{"p":"kilo","id":"z-ai/glm-4-32b","n":"Z.ai: GLM 4 32B ","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.1,"co":0.1,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-25","lu":"2026-03-15"},{"p":"kilo","id":"z-ai/glm-4.6v","n":"Z.ai: GLM 4.6V","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"ctx":131072,"out":131072,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-30","lu":"2026-01-10"},{"p":"kilo","id":"z-ai/glm-5v-turbo","n":"Z.ai: GLM 5V Turbo","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":1.2,"co":4,"cr":0.24,"ctx":202752,"out":131072,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-11"},{"p":"kilo","id":"z-ai/glm-4.5-air","n":"Z.ai: GLM 4.5 Air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.13,"co":0.85,"cr":0.025,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-28","lu":"2025-07-28"},{"p":"kilo","id":"z-ai/glm-4.7-flash","n":"Z.ai: GLM 4.7 Flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.06,"co":0.4,"cr":0.01,"ctx":202752,"out":40551,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"kilo","id":"z-ai/glm-5","n":"Z.ai: GLM 5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.72,"co":2.3,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-03-15"},{"p":"kilo","id":"z-ai/glm-5-turbo","n":"Z.ai: GLM 5 Turbo","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":1.2,"co":4,"cr":0.24,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-15","lu":"2026-04-11"},{"p":"kilo","id":"openai/gpt-4o-mini-2024-07-18","n":"OpenAI: GPT-4o-mini (2024-07-18)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2024-07-18","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-oss-safeguard-20b","n":"OpenAI: gpt-oss-safeguard-20b","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.075,"co":0.3,"cr":0.037,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-10-29","lu":"2025-10-29"},{"p":"kilo","id":"openai/gpt-3.5-turbo-instruct","n":"OpenAI: GPT-3.5 Turbo Instruct","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":4095,"out":4096,"mi":"t","mo":"t","rd":"2023-03-01","lu":"2023-09-21"},{"p":"kilo","id":"openai/gpt-5.2-chat","n":"OpenAI: GPT-5.2 Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2025-12-11","lu":"2026-03-15"},{"p":"kilo","id":"openai/o3","n":"OpenAI: o3","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2026-03-15"},{"p":"kilo","id":"openai/o4-mini-high","n":"OpenAI: o4 Mini High","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-04-17","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-audio","n":"OpenAI: GPT Audio","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"at","mo":"at","rd":"2026-01-20","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.2-pro","n":"OpenAI: GPT-5.2 Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-12-11","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4o-mini-search-preview","n":"OpenAI: GPT-4o-mini Search Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-01","lu":"2025-01"},{"p":"kilo","id":"openai/gpt-5","n":"OpenAI: GPT-5","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-08-07","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5-chat","n":"OpenAI: GPT-5 Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","rd":"2025-08-07","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-3.5-turbo","n":"OpenAI: GPT-3.5 Turbo","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"ctx":16385,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2023-03-01","lu":"2023-11-06"},{"p":"kilo","id":"openai/gpt-5-pro","n":"OpenAI: GPT-5 Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":15,"co":120,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-10-06","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4o","n":"OpenAI: GPT-4o","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2024-05-13","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4","n":"OpenAI: GPT-4","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8191,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2023-03-14","lu":"2024-04-09"},{"p":"kilo","id":"openai/o4-mini","n":"OpenAI: o4 Mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-3.5-turbo-16k","n":"OpenAI: GPT-3.5 Turbo 16k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":4,"ctx":16385,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2023-08-28","lu":"2026-03-15"},{"p":"kilo","id":"openai/o3-pro","n":"OpenAI: o3 Pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":20,"co":80,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.1-chat","n":"OpenAI: GPT-5.1 Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2025-11-13","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4o-2024-05-13","n":"OpenAI: GPT-4o (2024-05-13)","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":15,"ctx":128000,"out":4096,"mi":"ipt","mo":"t","tc":1,"rd":"2024-05-13","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4-0314","n":"OpenAI: GPT-4 (older v0314)","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8191,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2023-05-28","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.4-nano","n":"OpenAI: GPT-5.4 Nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-03-17","lu":"2026-04-11"},{"p":"kilo","id":"openai/gpt-5.3-chat","n":"OpenAI: GPT-5.3 Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2026-03-04","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-3.5-turbo-0613","n":"OpenAI: GPT-3.5 Turbo (older v0613)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":2,"ctx":4095,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2023-06-13","lu":"2023-06-13"},{"p":"kilo","id":"openai/gpt-5-image-mini","n":"OpenAI: GPT-5 Image Mini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2.5,"co":2,"ctx":400000,"out":128000,"mi":"ipt","mo":"it","rs":1,"tc":1,"rd":"2025-10-16","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.1-codex","n":"OpenAI: GPT-5.1-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"kilo","id":"openai/gpt-5.1-codex-max","n":"OpenAI: GPT-5.1-Codex-Max","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"kilo","id":"openai/gpt-4o-2024-08-06","n":"OpenAI: GPT-4o (2024-08-06)","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2024-08-06","lu":"2026-03-15"},{"p":"kilo","id":"openai/o3-mini","n":"OpenAI: o3 Mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"pt","mo":"t","tc":1,"rd":"2024-12-20","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.2","n":"OpenAI: GPT-5.2","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-12-11","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.3-codex","n":"OpenAI: GPT-5.3-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2026-02-25","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-audio-mini","n":"OpenAI: GPT Audio Mini","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0.6,"co":2.4,"ctx":128000,"out":16384,"mi":"at","mo":"at","rd":"2026-01-20","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.1-codex-mini","n":"OpenAI: GPT-5.1-Codex-Mini","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":100000,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"kilo","id":"openai/o4-mini-deep-research","n":"OpenAI: o4 Mini Deep Research","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2024-06-26","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4.1-nano","n":"OpenAI: GPT-4.1 Nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ipt","mo":"t","tc":1,"rd":"2025-04-14","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-oss-120b","n":"OpenAI: gpt-oss-120b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.039,"co":0.19,"ctx":131072,"out":26215,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"kilo","id":"openai/gpt-4o-2024-11-20","n":"OpenAI: GPT-4o (2024-11-20)","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2024-11-20","lu":"2026-03-15"},{"p":"kilo","id":"openai/o1","n":"OpenAI: o1","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","tc":1,"rd":"2024-12-05","lu":"2026-03-15"},{"p":"kilo","id":"openai/o1-pro","n":"OpenAI: o1-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":150,"co":600,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"rd":"2025-03-19","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-chat-latest","n":"OpenAI: GPT Chat Latest","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":5,"co":30,"cr":0.5,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-05-05","lu":"2026-05-07"},{"p":"kilo","id":"openai/gpt-5-image","n":"OpenAI: GPT-5 Image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":10,"co":10,"ctx":400000,"out":128000,"mi":"ipt","mo":"it","rs":1,"tc":1,"rd":"2025-10-14","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.4","n":"OpenAI: GPT-5.4","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"ctx":1050000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-03-06","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.4-mini","n":"OpenAI: GPT-5.4 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-03-17","lu":"2026-04-11"},{"p":"kilo","id":"openai/gpt-4.1","n":"OpenAI: GPT-4.1","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"ipt","mo":"t","tc":1,"rd":"2025-04-14","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4o-audio-preview","n":"OpenAI: GPT-4o Audio","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"at","mo":"at","tc":1,"rd":"2025-08-15","lu":"2026-03-15"},{"p":"kilo","id":"openai/o3-deep-research","n":"OpenAI: o3 Deep Research","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":10,"co":40,"cr":2.5,"ctx":200000,"out":100000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2024-06-26","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4-turbo-preview","n":"OpenAI: GPT-4 Turbo Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2024-01-25","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5-mini","n":"OpenAI: GPT-5 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-08-07","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4.1-mini","n":"OpenAI: GPT-4.1 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"ipt","mo":"t","tc":1,"rd":"2025-04-14","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-4-turbo","n":"OpenAI: GPT-4 Turbo","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2023-09-13","lu":"2024-04-09"},{"p":"kilo","id":"openai/gpt-5-nano","n":"OpenAI: GPT-5 Nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-08-07","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.4-pro","n":"OpenAI: GPT-5.4 Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-03-06","lu":"2026-03-15"},{"p":"kilo","id":"openai/o3-mini-high","n":"OpenAI: o3 Mini High","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"pt","mo":"t","tc":1,"rd":"2025-01-31","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.4-image-2","n":"OpenAI: GPT-5.4 Image 2","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":8,"co":15,"cr":2,"ctx":272000,"out":128000,"mi":"itp","mo":"it","rs":1,"rd":"2026-04-21","lu":"2026-05-01"},{"p":"kilo","id":"openai/gpt-4o-search-preview","n":"OpenAI: GPT-4o Search Preview","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-03-13","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.5-pro","n":"OpenAI: GPT-5.5 Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-05-01"},{"p":"kilo","id":"openai/gpt-4o-mini","n":"OpenAI: GPT-4o-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"ipt","mo":"t","tc":1,"rd":"2024-07-18","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-oss-20b","n":"OpenAI: gpt-oss-20b","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.03,"co":0.14,"ctx":131072,"out":26215,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"kilo","id":"openai/gpt-4-1106-preview","n":"OpenAI: GPT-4 Turbo (older v1106)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2023-11-06","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5-codex","n":"OpenAI: GPT-5 Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-09-15","lu":"2025-09-15"},{"p":"kilo","id":"openai/gpt-5.2-codex","n":"OpenAI: GPT-5.2-Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01-14","lu":"2026-01-14"},{"p":"kilo","id":"openai/gpt-5.1","n":"OpenAI: GPT-5.1","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-11-13","lu":"2026-03-15"},{"p":"kilo","id":"openai/gpt-5.5","n":"OpenAI: GPT-5.5","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-05-01"},{"p":"kilo","id":"thedrummer/cydonia-24b-v4.1","n":"TheDrummer: Cydonia 24B V4.1","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.3,"co":0.5,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"rd":"2025-09-27","lu":"2026-03-15"},{"p":"kilo","id":"thedrummer/skyfall-36b-v2","n":"TheDrummer: Skyfall 36B V2","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.55,"co":0.8,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2025-03-11","lu":"2026-03-15"},{"p":"kilo","id":"thedrummer/unslopnemo-12b","n":"TheDrummer: UnslopNemo 12B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.4,"co":0.4,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-09","lu":"2026-03-15"},{"p":"kilo","id":"thedrummer/rocinante-12b","n":"TheDrummer: Rocinante 12B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.17,"co":0.43,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09-30","lu":"2026-03-15"},{"p":"kilo","id":"bytedance/ui-tars-1.5-7b","n":"ByteDance: UI-TARS 7B ","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.1,"co":0.2,"ctx":128000,"out":2048,"mi":"it","mo":"t","rd":"2025-07-23","lu":"2026-03-15"},{"p":"kilo","id":"rekaai/reka-flash-3","n":"Reka Flash 3","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.1,"co":0.2,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-03-12","lu":"2026-04-11"},{"p":"kilo","id":"rekaai/reka-edge","n":"Reka Edge","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.1,"co":0.1,"ctx":16384,"out":16384,"mi":"itv","mo":"t","tc":1,"ow":1,"rd":"2026-03-20","lu":"2026-04-11"},{"p":"kilo","id":"mistralai/mistral-large-2407","n":"Mistral Large 2407","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-11-19","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/mistral-small-3.2-24b-instruct","n":"Mistral: Mistral Small 3.2 24B","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.06,"co":0.18,"cr":0.03,"ctx":131072,"out":131072,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2025-06-20","lu":"2025-06-20"},{"p":"kilo","id":"mistralai/mistral-nemo","n":"Mistral: Mistral Nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.02,"co":0.04,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-01","lu":"2024-07-30"},{"p":"kilo","id":"mistralai/mistral-medium-3-5","n":"Mistral: Mistral Medium 3.5","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":1.5,"co":7.5,"ctx":262144,"out":262144,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2026-04-30","lu":"2026-05-07"},{"p":"kilo","id":"mistralai/ministral-8b-2512","n":"Mistral: Ministral 3 8B 2512","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.15,"co":0.15,"ctx":262144,"out":32768,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/devstral-small","n":"Mistral: Devstral Small 1.1","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.1,"co":0.3,"ctx":131072,"out":26215,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-05-07","lu":"2025-07-10"},{"p":"kilo","id":"mistralai/mistral-small-3.1-24b-instruct","n":"Mistral: Mistral Small 3.1 24B","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.35,"co":0.56,"cr":0.015,"ctx":128000,"out":131072,"mi":"it","mo":"t","ow":1,"rd":"2025-03-17","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/mistral-saba","n":"Mistral: Saba","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.2,"co":0.6,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-02-17","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/mistral-large","n":"Mistral Large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":128000,"out":25600,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-24","lu":"2025-12-02"},{"p":"kilo","id":"mistralai/mistral-medium-3.1","n":"Mistral: Mistral Medium 3.1","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":131072,"out":26215,"mi":"ti","mo":"t","tc":1,"rd":"2025-08-12","lu":"2025-08-12"},{"p":"kilo","id":"mistralai/pixtral-large-2411","n":"Mistral: Pixtral Large 2411","d":"Mistral vision-language model for image understanding and multimodal chat","ci":2,"co":6,"ctx":131072,"out":32768,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2024-11-19","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/devstral-medium","n":"Mistral: Devstral Medium","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":2,"ctx":131072,"out":26215,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-10","lu":"2025-07-10"},{"p":"kilo","id":"mistralai/mistral-small-24b-instruct-2501","n":"Mistral: Mistral Small 3","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.05,"co":0.08,"ctx":32768,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-29","lu":"2026-01-10"},{"p":"kilo","id":"mistralai/ministral-3b-2512","n":"Mistral: Ministral 3 3B 2512","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"ctx":131072,"out":32768,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/mistral-small-2603","n":"Mistral: Mistral Small 4","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.15,"co":0.6,"cr":0.015,"ctx":262144,"out":262144,"mi":"it","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-16","lu":"2026-04-11"},{"p":"kilo","id":"mistralai/mistral-large-2411","n":"Mistral Large 2411","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":131072,"out":26215,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-24","lu":"2024-11-04"},{"p":"kilo","id":"mistralai/mistral-7b-instruct-v0.1","n":"Mistral: Mistral 7B Instruct v0.1","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.11,"co":0.19,"ctx":2824,"out":565,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"kilo","id":"mistralai/ministral-14b-2512","n":"Mistral: Ministral 3 14B 2512","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"ctx":262144,"out":52429,"mi":"ti","mo":"t","tc":1,"rd":"2025-12-16","lu":"2025-12-16"},{"p":"kilo","id":"mistralai/devstral-2512","n":"Mistral: Devstral 2 2512","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":2,"cr":0.025,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-09-12","lu":"2026-03-15"},{"p":"kilo","id":"mistralai/mixtral-8x22b-instruct","n":"Mistral: Mixtral 8x22B Instruct","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":2,"co":6,"ctx":65536,"out":13108,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-04-17","lu":"2024-04-17"},{"p":"kilo","id":"mistralai/mistral-medium-3","n":"Mistral: Mistral Medium 3","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":131072,"out":26215,"mi":"ti","mo":"t","tc":1,"rd":"2025-05-07","lu":"2025-05-07"},{"p":"kilo","id":"mistralai/voxtral-small-24b-2507","n":"Mistral: Voxtral Small 24B 2507","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"ctx":32000,"out":6400,"mi":"ta","mo":"t","tc":1,"ow":1,"rd":"2025-07-01","lu":"2025-07-01"},{"p":"kilo","id":"mistralai/mistral-large-2512","n":"Mistral: Mistral Large 3 2512","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0.5,"co":1.5,"ctx":262144,"out":52429,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2024-11-01","lu":"2025-12-16"},{"p":"kilo","id":"mistralai/codestral-2508","n":"Mistral: Codestral 2508","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.9,"ctx":256000,"out":51200,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-08-01","lu":"2025-08-01"},{"p":"kilo","id":"morph/morph-v3-fast","n":"Morph: Morph V3 Fast","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.8,"co":1.2,"ctx":81920,"out":38000,"mi":"t","mo":"t","rd":"2024-08-15","lu":"2024-08-15"},{"p":"kilo","id":"morph/morph-v3-large","n":"Morph: Morph V3 Large","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.9,"co":1.9,"ctx":262144,"out":131072,"mi":"t","mo":"t","rd":"2024-08-15","lu":"2024-08-15"},{"p":"kilo","id":"bytedance-seed/seed-1.6-flash","n":"ByteDance Seed: Seed 1.6 Flash","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.075,"co":0.3,"ctx":262144,"out":32768,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2026-03-15"},{"p":"kilo","id":"bytedance-seed/seed-1.6","n":"ByteDance Seed: Seed 1.6","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"ctx":262144,"out":32768,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2025-09","lu":"2025-09"},{"p":"kilo","id":"bytedance-seed/seed-2.0-mini","n":"ByteDance Seed: Seed-2.0-Mini","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.1,"co":0.4,"ctx":262144,"out":131072,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-27","lu":"2026-03-15"},{"p":"kilo","id":"bytedance-seed/seed-2.0-lite","n":"ByteDance Seed: Seed-2.0-Lite","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"ctx":262144,"out":131072,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-10","lu":"2026-03-15"},{"p":"kilo","id":"anthracite-org/magnum-v4-72b","n":"Magnum v4 72B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":3,"co":5,"ctx":16384,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2024-10-22","lu":"2026-03-15"},{"p":"kilo","id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","n":"NVIDIA: Nemotron 3 Nano Omni (free)","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0,"co":0,"ctx":256000,"out":65536,"mi":"taiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-05-01"},{"p":"kilo","id":"nvidia/nemotron-3-nano-30b-a3b","n":"NVIDIA: Nemotron 3 Nano 30B A3B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.05,"co":0.2,"ctx":262144,"out":52429,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2024-12","lu":"2026-02-04"},{"p":"kilo","id":"nvidia/llama-3.3-nemotron-super-49b-v1.5","n":"NVIDIA: Llama 3.3 Nemotron Super 49B V1.5","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.1,"co":0.4,"ctx":131072,"out":26215,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-03-16","lu":"2025-03-16"},{"p":"kilo","id":"nvidia/nemotron-nano-9b-v2","n":"NVIDIA: Nemotron Nano 9B V2","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0.04,"co":0.16,"ctx":131072,"out":26215,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-18","lu":"2025-08-18"},{"p":"kilo","id":"nvidia/nemotron-3-super-120b-a12b:free","n":"NVIDIA: Nemotron 3 Super (free)","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-12","lu":"2026-03-15"},{"p":"kilo","id":"nvidia/nemotron-3-super-120b-a12b","n":"NVIDIA: Nemotron 3 Super","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.1,"co":0.5,"cr":0.1,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-04-11"},{"p":"kilo","id":"xiaomi/mimo-v2.5","n":"Xiaomi: MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.4,"co":2,"cr":0.08,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"kilo","id":"xiaomi/mimo-v2-omni","n":"Xiaomi: MiMo-V2-Omni","f":"mimo","d":"MiMo omni model for text, image, video, audio, and agents","ci":0.4,"co":2,"cr":0.08,"ctx":262144,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"kilo","id":"xiaomi/mimo-v2-flash","n":"Xiaomi: MiMo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.09,"co":0.29,"cr":0.045,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12-01","rd":"2025-12-16","lu":"2026-02-04"},{"p":"kilo","id":"xiaomi/mimo-v2-pro","n":"Xiaomi: MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":1,"co":3,"cr":0.2,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"kilo","id":"xiaomi/mimo-v2.5-pro","n":"Xiaomi: MiMo V2.5 Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":1,"co":3,"cr":0.2,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"kilo","id":"inception/mercury-2","n":"Inception: Mercury 2","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.25,"co":0.75,"cr":0.025,"ctx":128000,"out":50000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"kilo","id":"anthropic/claude-3.5-haiku","n":"Anthropic: Claude 3.5 Haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2024-10-22","lu":"2024-10-22"},{"p":"kilo","id":"anthropic/claude-sonnet-4.5","n":"Anthropic: Claude Sonnet 4.5","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2026-03-15"},{"p":"kilo","id":"anthropic/claude-sonnet-4","n":"Anthropic: Claude Sonnet 4","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2026-03-15"},{"p":"kilo","id":"anthropic/claude-opus-4.6-fast","n":"Anthropic: Claude Opus 4.6 (Fast)","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":30,"co":150,"cr":3,"cw":37.5,"ctx":1000000,"out":128000,"mi":"it","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-04-07","lu":"2026-04-11"},{"p":"kilo","id":"anthropic/claude-haiku-4.5","n":"Anthropic: Claude Haiku 4.5","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"kilo","id":"anthropic/claude-opus-4.7-fast","n":"Anthropic: Claude Opus 4.7 (Fast)","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":30,"co":150,"cr":3,"cw":37.5,"ctx":1000000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-16"},{"p":"kilo","id":"anthropic/claude-opus-4.7","n":"Anthropic: Claude Opus 4.7","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-05-01"},{"p":"kilo","id":"anthropic/claude-opus-4.1","n":"Anthropic: Claude Opus 4.1","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2026-03-15"},{"p":"kilo","id":"anthropic/claude-opus-4.5","n":"Anthropic: Claude Opus 4.5","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-11-24","lu":"2026-03-15"},{"p":"kilo","id":"anthropic/claude-sonnet-4.6","n":"Anthropic: Claude Sonnet 4.6","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"ctx":1000000,"out":128000,"mi":"it","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-15"},{"p":"kilo","id":"anthropic/claude-3-haiku","n":"Anthropic: Claude 3 Haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.25,"co":1.25,"cr":0.03,"cw":0.3,"ctx":200000,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-03-07","lu":"2024-03-07"},{"p":"kilo","id":"anthropic/claude-opus-4","n":"Anthropic: Claude Opus 4","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2026-03-15"},{"p":"kilo","id":"anthropic/claude-opus-4.6","n":"Anthropic: Claude Opus 4.6","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"kilo","id":"tencent/hunyuan-a13b-instruct","n":"Tencent: Hunyuan A13B Instruct","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.14,"co":0.57,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"rd":"2025-06-30","lu":"2025-11-25"},{"p":"kilo","id":"tencent/hy3-preview","n":"Tencent: Hy3 Preview","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.066,"co":0.26,"cr":0.029,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-22","lu":"2026-05-16"},{"p":"kilo","id":"deepcogito/cogito-v2.1-671b","n":"Deep Cogito: Cogito v2.1 671B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1.25,"co":1.25,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-11-14","lu":"2026-03-15"},{"p":"kilo","id":"cohere/command-a","n":"Cohere: Command A","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2025-03-13","lu":"2025-03-13"},{"p":"kilo","id":"cohere/command-r-08-2024","n":"Cohere: Command R (08-2024)","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.15,"co":0.6,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-08-30","lu":"2024-08-30"},{"p":"kilo","id":"cohere/command-r7b-12-2024","n":"Cohere: Command R7B (12-2024)","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":0.0375,"co":0.15,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-12-02","lu":"2024-12-02"},{"p":"kilo","id":"cohere/command-r-plus-08-2024","n":"Cohere: Command R+ (08-2024)","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":2.5,"co":10,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-08-30","lu":"2024-08-30"},{"p":"kilo","id":"gryphe/mythomax-l2-13b","n":"MythoMax 13B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.06,"co":0.06,"ctx":4096,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2024-04-25","lu":"2024-04-25"},{"p":"kilo","id":"stepfun/step-3.5-flash","n":"StepFun: Step 3.5 Flash","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"cr":0.02,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-29","lu":"2026-01-29"},{"p":"kilo","id":"prime-intellect/intellect-3","n":"Prime Intellect: INTELLECT-3","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.2,"co":1.1,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-11-26","lu":"2026-02-04"},{"p":"kilo","id":"nex-agi/deepseek-v3.1-nex-n1","n":"Nex AGI: DeepSeek V3.1 Nex N1","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":131072,"out":163840,"mi":"t","mo":"t","tc":1,"rd":"2025-01-01","lu":"2025-11-25"},{"p":"kilo","id":"undi95/remm-slerp-l2-13b","n":"ReMM SLERP 13B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.45,"co":0.65,"ctx":6144,"out":4096,"mi":"t","mo":"t","ow":1,"rd":"2023-07-22","lu":"2026-03-15"},{"p":"kilo","id":"~openai/gpt-mini-latest","n":"OpenAI: GPT Mini Latest","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"~openai/gpt-latest","n":"OpenAI: GPT Latest","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"~moonshotai/kimi-latest","n":"MoonshotAI: Kimi Latest","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.74,"co":3.49,"cr":0.14,"ctx":262142,"out":262142,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"relace/relace-search","n":"Relace: Relace Search","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":1,"co":3,"ctx":256000,"out":128000,"mi":"t","mo":"t","tc":1,"rd":"2025-12-09","lu":"2026-03-15"},{"p":"kilo","id":"relace/relace-apply-3","n":"Relace: Relace Apply 3","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.85,"co":1.25,"ctx":256000,"out":128000,"mi":"t","mo":"t","rd":"2025-09-26","lu":"2026-03-15"},{"p":"kilo","id":"ai21/jamba-large-1.7","n":"AI21: Jamba Large 1.7","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":2,"co":8,"ctx":256000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2025-08-09","lu":"2026-03-15"},{"p":"kilo","id":"arcee-ai/coder-large","n":"Arcee AI: Coder Large","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.5,"co":0.8,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2025-05-06","lu":"2026-03-15"},{"p":"kilo","id":"arcee-ai/virtuoso-large","n":"Arcee AI: Virtuoso Large","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.75,"co":1.2,"ctx":131072,"out":64000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-05-06","lu":"2026-03-15"},{"p":"kilo","id":"arcee-ai/spotlight","n":"Arcee AI: Spotlight","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.18,"co":0.18,"ctx":131072,"out":65537,"mi":"it","mo":"t","ow":1,"rd":"2025-05-06","lu":"2026-03-15"},{"p":"kilo","id":"arcee-ai/trinity-large-thinking","n":"Arcee AI: Trinity Large Thinking","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.22,"co":0.85,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-11"},{"p":"kilo","id":"arcee-ai/maestro-reasoning","n":"Arcee AI: Maestro Reasoning","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.9,"co":3.3,"ctx":131072,"out":32000,"mi":"t","mo":"t","ow":1,"rd":"2025-05-06","lu":"2026-03-15"},{"p":"kilo","id":"arcee-ai/trinity-mini","n":"Arcee AI: Trinity Mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.045,"co":0.15,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12","lu":"2026-01-28"},{"p":"kilo","id":"mancer/weaver","n":"Mancer: Weaver (alpha)","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.75,"co":1,"ctx":8000,"out":2000,"mi":"t","mo":"t","rd":"2023-08-02","lu":"2026-03-15"},{"p":"kilo","id":"perplexity/sonar-reasoning-pro","n":"Perplexity: Sonar Reasoning Pro","d":"Web-grounded reasoning model for multi-step research and cited answers","ci":2,"co":8,"ctx":128000,"out":25600,"mi":"ti","mo":"t","rs":1,"rd":"2024-01-01","lu":"2025-09-01"},{"p":"kilo","id":"perplexity/sonar","n":"Perplexity: Sonar","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":1,"co":1,"ctx":127072,"out":25415,"mi":"ti","mo":"t","rd":"2024-01-01","lu":"2025-09-01"},{"p":"kilo","id":"perplexity/sonar-pro","n":"Perplexity: Sonar Pro","d":"Advanced Sonar search model for deeper research and cited synthesis","ci":3,"co":15,"ctx":200000,"out":8000,"mi":"ti","mo":"t","rd":"2024-01-01","lu":"2025-09-01"},{"p":"kilo","id":"perplexity/sonar-pro-search","n":"Perplexity: Sonar Pro Search","d":"Advanced Sonar search model for deeper research and cited synthesis","ci":3,"co":15,"ctx":200000,"out":8000,"mi":"it","mo":"t","rs":1,"rd":"2025-10-31","lu":"2026-03-15"},{"p":"kilo","id":"perplexity/sonar-deep-research","n":"Perplexity: Sonar Deep Research","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":2,"co":8,"ctx":128000,"out":25600,"mi":"t","mo":"t","rs":1,"rd":"2025-01-27","lu":"2025-01-27"},{"p":"kilo","id":"switchpoint/router","n":"Switchpoint Router","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.85,"co":3.4,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-07-12","lu":"2026-03-15"},{"p":"kilo","id":"openrouter/bodybuilder","n":"Body Builder (beta)","d":"Preview model for early access evaluation, prototyping, and compatibility testing","ci":0,"co":0,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2026-03-15","lu":"2026-03-15"},{"p":"kilo","id":"openrouter/free","n":"Free Models Router","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0,"co":0,"ctx":200000,"out":32768,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2026-02-01","lu":"2026-03-15"},{"p":"kilo","id":"openrouter/owl-alpha","n":"Owl Alpha","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0,"co":0,"ctx":1048756,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-04-30"},{"p":"kilo","id":"openrouter/pareto-code","n":"Pareto Code Router","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0,"co":0,"ctx":200000,"out":65536,"mi":"t","mo":"t","rd":"2026-04-21","lu":"2026-05-01"},{"p":"kilo","id":"openrouter/auto","n":"Auto Router","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":2000000,"out":32768,"mi":"aiptv","mo":"it","rs":1,"tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.5-plus-20260420","n":"Qwen: Qwen3.5 Plus 2026-04-20","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":2.4,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"qwen/qwen3-vl-235b-a22b-thinking","n":"Qwen: Qwen3 VL 235B A22B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.6,"ctx":131072,"out":32768,"mi":"it","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-24","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-vl-30b-a3b-thinking","n":"Qwen: Qwen3 VL 30B A3B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.13,"co":1.56,"ctx":131072,"out":32768,"mi":"it","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-11","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-coder-plus","n":"Qwen: Qwen3 Coder Plus","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.65,"co":3.25,"cr":0.2,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-01","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen-plus","n":"Qwen: Qwen-Plus","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.4,"co":1.2,"cr":0.08,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2024-01-25","lu":"2025-09-11"},{"p":"kilo","id":"qwen/qwen3-coder-30b-a3b-instruct","n":"Qwen: Qwen3 Coder 30B A3B Instruct","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.07,"co":0.27,"ctx":160000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-31","lu":"2025-07-31"},{"p":"kilo","id":"qwen/qwen3-32b","n":"Qwen: Qwen3 32B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.08,"co":0.24,"cr":0.04,"ctx":40960,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2024-12-01","lu":"2026-02-04"},{"p":"kilo","id":"qwen/qwen3-next-80b-a3b-instruct","n":"Qwen: Qwen3 Next 80B A3B Instruct","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":1.1,"ctx":131072,"out":52429,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-09-11","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-vl-8b-instruct","n":"Qwen: Qwen3 VL 8B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.08,"co":0.5,"ctx":131072,"out":32768,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2025-10-15","lu":"2025-11-25"},{"p":"kilo","id":"qwen/qwen3.6-35b-a3b","n":"Qwen: Qwen3.6 35B A3B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1612,"co":0.96525,"cr":0.1612,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"qwen/qwen3.7-max","n":"Qwen: Qwen3.7 Max","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.625,"co":4.875,"cr":0.1625,"cw":2.03125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-26","lu":"2026-05-27"},{"p":"kilo","id":"qwen/qwen3-max","n":"Qwen: Qwen3 Max","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.2,"co":6,"cr":0.24,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2025-09-05","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-8b","n":"Qwen: Qwen3 8B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.05,"co":0.4,"cr":0.05,"ctx":40960,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen-plus-2025-07-28","n":"Qwen: Qwen Plus 0728","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.26,"co":0.78,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-09-09","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.5-flash-02-23","n":"Qwen: Qwen3.5-Flash","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1,"co":0.4,"ctx":1000000,"out":65536,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-30b-a3b-instruct-2507","n":"Qwen: Qwen3 30B A3B Instruct 2507","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.3,"cr":0.04,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-29","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen-2.5-coder-32b-instruct","n":"Qwen2.5 Coder 32B Instruct","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.2,"co":0.2,"cr":0.015,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-11-11","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-next-80b-a3b-thinking","n":"Qwen: Qwen3 Next 80B A3B Thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.0975,"co":0.78,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-11","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-235b-a22b-thinking-2507","n":"Qwen: Qwen3 235B A22B Thinking 2507","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.11,"co":0.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-25","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-vl-32b-instruct","n":"Qwen: Qwen3 VL 32B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.104,"co":0.416,"ctx":131072,"out":32768,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-21","lu":"2025-11-25"},{"p":"kilo","id":"qwen/qwen3-coder","n":"Qwen: Qwen3 Coder 480B A35B","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.22,"co":1,"cr":0.022,"ctx":262144,"out":52429,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-23","lu":"2025-07-23"},{"p":"kilo","id":"qwen/qwen3.6-flash","n":"Qwen: Qwen3.6 Flash","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.25,"co":1.5,"cw":0.3125,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"qwen/qwen3.5-plus-02-15","n":"Qwen: Qwen3.5 Plus 2026-02-15","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":1.56,"ctx":1000000,"out":65536,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2026-02-15","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen-2.5-7b-instruct","n":"Qwen: Qwen2.5 7B Instruct","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.04,"co":0.1,"ctx":32768,"out":6554,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09","lu":"2025-04-16"},{"p":"kilo","id":"qwen/qwen3-vl-8b-thinking","n":"Qwen: Qwen3 VL 8B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.117,"co":1.365,"ctx":131072,"out":32768,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2025-10-15","lu":"2025-11-25"},{"p":"kilo","id":"qwen/qwen3-max-thinking","n":"Qwen: Qwen3 Max Thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.78,"co":3.9,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-01-23","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-30b-a3b-thinking-2507","n":"Qwen: Qwen3 30B A3B Thinking 2507","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.051,"co":0.34,"ctx":32768,"out":6554,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-29","lu":"2025-07-29"},{"p":"kilo","id":"qwen/qwen2.5-vl-72b-instruct","n":"Qwen: Qwen2.5 VL 72B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.8,"co":0.8,"cr":0.075,"ctx":32768,"out":32768,"mi":"it","mo":"t","ow":1,"rd":"2025-02-01","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.5-27b","n":"Qwen: Qwen3.5-27B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.195,"co":1.56,"ctx":262144,"out":65536,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-235b-a22b","n":"Qwen: Qwen3 235B A22B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.455,"co":1.82,"cr":0.15,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2024-12-01","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen-2.5-72b-instruct","n":"Qwen2.5 72B Instruct","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.12,"co":0.39,"ctx":32768,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09","lu":"2026-01-10"},{"p":"kilo","id":"qwen/qwen-plus-2025-07-28:thinking","n":"Qwen: Qwen Plus 0728 (thinking)","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.26,"co":0.78,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-09","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-coder-next","n":"Qwen: Qwen3 Coder Next","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.12,"co":0.75,"cr":0.035,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-02","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.6-27b","n":"Qwen: Qwen3.6 27B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.325,"co":3.25,"ctx":256000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"qwen/qwen3.5-35b-a3b","n":"Qwen: Qwen3.5-35B-A3B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1625,"co":1.3,"ctx":262144,"out":65536,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.5-9b","n":"Qwen: Qwen3.5-9B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.05,"co":0.15,"ctx":256000,"out":32768,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-10","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.5-397b-a17b","n":"Qwen: Qwen3.5 397B A17B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.39,"co":2.34,"ctx":262144,"out":65536,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2026-02-15","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-vl-30b-a3b-instruct","n":"Qwen: Qwen3 VL 30B A3B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.13,"co":0.52,"ctx":131072,"out":32768,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-10-05","lu":"2025-11-25"},{"p":"kilo","id":"qwen/qwen3-235b-a22b-2507","n":"Qwen: Qwen3 235B A22B Instruct 2507","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.071,"co":0.1,"ctx":262144,"out":52429,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-04","lu":"2026-01"},{"p":"kilo","id":"qwen/qwen3-coder-flash","n":"Qwen: Qwen3 Coder Flash","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.195,"co":0.975,"cr":0.06,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-07-23","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-14b","n":"Qwen: Qwen3 14B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.06,"co":0.24,"cr":0.025,"ctx":40960,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3-vl-235b-a22b-instruct","n":"Qwen: Qwen3 VL 235B A22B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.88,"cr":0.11,"ctx":262144,"out":52429,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-09-23","lu":"2026-01-10"},{"p":"kilo","id":"qwen/qwen3-30b-a3b","n":"Qwen: Qwen3 30B A3B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.08,"co":0.28,"cr":0.03,"ctx":40960,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-04","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.6-max-preview","n":"Qwen: Qwen3.6 Max Preview","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.04,"co":6.24,"cw":1.3,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-05-01"},{"p":"kilo","id":"qwen/qwen3.5-122b-a10b","n":"Qwen: Qwen3.5-122B-A10B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.08,"ctx":262144,"out":65536,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-26","lu":"2026-03-15"},{"p":"kilo","id":"qwen/qwen3.6-plus","n":"Qwen: Qwen3.6 Plus","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.325,"co":1.95,"cr":0.0325,"cw":0.40625,"ctx":1000000,"out":65536,"mi":"it","mo":"t","rs":1,"tc":1,"rd":"2025-08-26","lu":"2026-04-11"},{"p":"kilo","id":"amazon/nova-lite-v1","n":"Amazon: Nova Lite 1.0","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.06,"co":0.24,"ctx":300000,"out":5120,"mi":"it","mo":"t","tc":1,"rd":"2024-12-06","lu":"2026-03-15"},{"p":"kilo","id":"amazon/nova-premier-v1","n":"Amazon: Nova Premier 1.0","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":2.5,"co":12.5,"ctx":1000000,"out":32000,"mi":"it","mo":"t","tc":1,"rd":"2025-11-01","lu":"2026-03-15"},{"p":"kilo","id":"amazon/nova-pro-v1","n":"Amazon: Nova Pro 1.0","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.8,"co":3.2,"ctx":300000,"out":5120,"mi":"ti","mo":"t","tc":1,"rd":"2024-12-03","lu":"2024-12-03"},{"p":"kilo","id":"amazon/nova-micro-v1","n":"Amazon: Nova Micro 1.0","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.035,"co":0.14,"ctx":128000,"out":5120,"mi":"t","mo":"t","tc":1,"rd":"2024-12-06","lu":"2026-03-15"},{"p":"kilo","id":"amazon/nova-2-lite-v1","n":"Amazon: Nova 2 Lite","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.3,"co":2.5,"ctx":1000000,"out":65535,"mi":"iptv","mo":"t","rs":1,"tc":1,"rd":"2024-12-01","lu":"2026-03-15"},{"p":"kilo","id":"aion-labs/aion-rp-llama-3.1-8b","n":"AionLabs: Aion-RP 1.0 (8B)","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.8,"co":1.6,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-02-05","lu":"2026-03-15"},{"p":"kilo","id":"aion-labs/aion-1.0-mini","n":"AionLabs: Aion-1.0-Mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.7,"co":1.4,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-02-05","lu":"2026-03-15"},{"p":"kilo","id":"aion-labs/aion-2.0","n":"AionLabs: Aion-2.0","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.8,"co":1.6,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2026-02-24","lu":"2026-03-15"},{"p":"kilo","id":"aion-labs/aion-1.0","n":"AionLabs: Aion-1.0","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":4,"co":8,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-02-05","lu":"2026-03-15"},{"p":"kilo","id":"inflection/inflection-3-pi","n":"Inflection: Inflection 3 Pi","d":"General-purpose chat model for instruction following, writing, and analysis","ci":2.5,"co":10,"ctx":8000,"out":1024,"mi":"t","mo":"t","rd":"2024-10-11","lu":"2026-03-15"},{"p":"kilo","id":"inflection/inflection-3-productivity","n":"Inflection: Inflection 3 Productivity","d":"General-purpose chat model for instruction following, writing, and analysis","ci":2.5,"co":10,"ctx":8000,"out":1024,"mi":"t","mo":"t","rd":"2024-10-11","lu":"2026-03-15"},{"p":"kilo","id":"sao10k/l3.1-euryale-70b","n":"Sao10K: Llama 3.1 Euryale 70B v2.2","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.85,"co":0.85,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-08-28","lu":"2026-03-15"},{"p":"kilo","id":"sao10k/l3.3-euryale-70b","n":"Sao10K: Llama 3.3 Euryale 70B","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.65,"co":0.75,"ctx":131072,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2024-12-18","lu":"2026-03-15"},{"p":"kilo","id":"sao10k/l3-lunaris-8b","n":"Sao10K: Llama 3 8B Lunaris","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.04,"co":0.05,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-08-13","lu":"2026-03-15"},{"p":"kilo","id":"sao10k/l3-euryale-70b","n":"Sao10k: Llama 3 Euryale 70B v2.1","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":1.48,"co":1.48,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-06-18","lu":"2026-03-15"},{"p":"kilo","id":"sao10k/l3.1-70b-hanami-x1","n":"Sao10K: Llama 3.1 70B Hanami x1","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":3,"co":3,"ctx":16000,"out":16000,"mi":"t","mo":"t","ow":1,"rd":"2025-01-08","lu":"2026-03-15"},{"p":"kilo","id":"upstage/solar-pro-3","n":"Upstage: Solar Pro 3","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.15,"co":0.6,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-01-27","lu":"2026-03-15"},{"p":"kilo","id":"allenai/olmo-3-32b-think","n":"AllenAI: Olmo 3 32B Think","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.15,"co":0.5,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-11-22","lu":"2026-03-15"},{"p":"kilo","id":"essentialai/rnj-1-instruct","n":"EssentialAI: Rnj 1 Instruct","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.15,"co":0.15,"ctx":32768,"out":6554,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-05","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-r1-0528","n":"DeepSeek: R1 0528","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.45,"co":2.15,"cr":0.2,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-05-28","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-v4-flash","n":"DeepSeek: DeepSeek V4 Flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-05-01"},{"p":"kilo","id":"deepseek/deepseek-v3.1-terminus","n":"DeepSeek: DeepSeek V3.1 Terminus","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.21,"co":0.79,"cr":0.13,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-22","lu":"2025-09-22"},{"p":"kilo","id":"deepseek/deepseek-r1-distill-llama-70b","n":"DeepSeek: R1 Distill Llama 70B","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.7,"co":0.8,"cr":0.015,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-01-23","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-v4-pro","n":"DeepSeek: DeepSeek V4 Pro","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-05-01"},{"p":"kilo","id":"deepseek/deepseek-r1","n":"DeepSeek: R1","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.7,"co":2.5,"ctx":64000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"kilo","id":"deepseek/deepseek-v3.2-speciale","n":"DeepSeek: DeepSeek V3.2 Speciale","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.4,"co":1.2,"cr":0.135,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-12-01","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-v3.2-exp","n":"DeepSeek: DeepSeek V3.2 Exp","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.41,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-01-01","lu":"2025-09-29"},{"p":"kilo","id":"deepseek/deepseek-chat-v3-0324","n":"DeepSeek: DeepSeek V3 0324","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2,"co":0.77,"cr":0.095,"ctx":163840,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-03-24","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-r1-distill-qwen-32b","n":"DeepSeek: R1 Distill Qwen 32B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.29,"co":0.29,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-01-01","lu":"2025-11-25"},{"p":"kilo","id":"deepseek/deepseek-chat","n":"DeepSeek: DeepSeek V3","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.32,"co":0.89,"cr":0.15,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-12-01","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-v3.2","n":"DeepSeek: DeepSeek V3.2","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.26,"co":0.38,"cr":0.125,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-01","lu":"2026-03-15"},{"p":"kilo","id":"deepseek/deepseek-chat-v3.1","n":"DeepSeek: DeepSeek V3.1","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.15,"co":0.75,"ctx":32768,"out":7168,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-21","lu":"2025-08-21"},{"p":"kilo","id":"minimax/minimax-m2.5","n":"MiniMax: MiniMax M2.5","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.25,"co":1.2,"cr":0.029,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-03-15"},{"p":"kilo","id":"minimax/minimax-m2.1","n":"MiniMax: MiniMax M2.1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.27,"co":0.95,"cr":0.03,"ctx":196608,"out":39322,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"kilo","id":"minimax/minimax-01","n":"MiniMax: MiniMax-01","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.2,"co":1.1,"ctx":1000192,"out":1000192,"mi":"ti","mo":"t","ow":1,"rd":"2025-01-15","lu":"2025-01-15"},{"p":"kilo","id":"minimax/minimax-m2","n":"MiniMax: MiniMax M2","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.255,"co":1,"cr":0.03,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-23","lu":"2026-03-15"},{"p":"kilo","id":"minimax/minimax-m2-her","n":"MiniMax: MiniMax M2-her","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":65536,"out":2048,"mi":"t","mo":"t","ow":1,"rd":"2026-01-23","lu":"2026-03-15"},{"p":"kilo","id":"minimax/minimax-m2.7","n":"MiniMax: MiniMax M2.7","f":"minimax-m2.7","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"kilo","id":"minimax/minimax-m1","n":"MiniMax: MiniMax M1","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.4,"co":2.2,"ctx":1000000,"out":40000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-06-17","lu":"2025-06-17"},{"p":"kilo","id":"stealth/claude-opus-4.7","n":"Stealth: Claude Opus 4.7 (20% off)","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4,"co":20,"cr":0.4,"cw":5,"ctx":1000000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-05-27"},{"p":"kilo","id":"stealth/claude-sonnet-4.6","n":"Stealth: Claude Sonnet 4.6 (20% off)","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.4,"co":12,"cr":0.24,"cw":3,"ctx":1000000,"out":64000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-02-17","lu":"2026-05-27"},{"p":"kilo","id":"stealth/claude-opus-4.6","n":"Stealth: Claude Opus 4.6 (20% off)","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4,"co":20,"cr":0.4,"cw":5,"ctx":1000000,"out":128000,"mi":"ipt","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-05-27"},{"p":"kilo","id":"kwaipilot/kat-coder-pro-v2","n":"Kwaipilot: KAT-Coder-Pro V2","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":256000,"out":80000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-04-11"},{"p":"kilo","id":"nousresearch/hermes-2-pro-llama-3-8b","n":"NousResearch: Hermes 2 Pro - Llama-3 8B","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.14,"co":0.14,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-05-27","lu":"2024-06-27"},{"p":"kilo","id":"nousresearch/hermes-4-405b","n":"Nous: Hermes 4 405B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1,"co":3,"ctx":131072,"out":26215,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-08-25","lu":"2025-08-25"},{"p":"kilo","id":"nousresearch/hermes-3-llama-3.1-70b","n":"Nous: Hermes 3 70B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.3,"co":0.3,"ctx":131072,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2024-08-18","lu":"2026-03-15"},{"p":"kilo","id":"nousresearch/hermes-3-llama-3.1-405b","n":"Nous: Hermes 3 405B Instruct","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":1,"co":1,"ctx":131072,"out":16384,"mi":"t","mo":"t","ow":1,"rd":"2024-08-16","lu":"2024-08-16"},{"p":"kilo","id":"nousresearch/hermes-4-70b","n":"Nous: Hermes 4 70B","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.13,"co":0.4,"cr":0.055,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-08-25","lu":"2026-03-15"},{"p":"lucidquery","id":"lucidnova-rf1-100b","n":"LucidNova RF1 100B","f":"nova","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":2,"co":5,"ctx":120000,"out":8000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-09-16","rd":"2024-12-28","lu":"2025-09-10"},{"p":"lucidquery","id":"lucidquery-nexus-coder","n":"LucidQuery Nexus Coder","f":"lucid","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":2,"co":5,"ctx":250000,"out":60000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-08-01","rd":"2025-09-01","lu":"2025-09-01"},{"p":"lucidquery","id":"lucidquery-agi-01-swift","n":"AGI-01 Swift","f":"agi","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":2.5,"co":15,"ctx":300000,"out":120000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-06-05","rd":"2026-06-16","lu":"2026-06-16"},{"p":"lucidquery","id":"lucidquery-agi-01-frontier","n":"AGI-01 Frontier","f":"agi","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":4.5,"co":22,"ctx":300000,"out":120000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-06-05","rd":"2026-06-16","lu":"2026-06-16"},{"p":"meganova","id":"meta-llama/Llama-3.3-70B-Instruct","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.1,"co":0.3,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-12-06","lu":"2024-12-06"},{"p":"meganova","id":"moonshotai/Kimi-K2-Thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"meganova","id":"moonshotai/Kimi-K2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.45,"co":2.8,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"meganova","id":"Qwen/Qwen2.5-VL-32B-Instruct","n":"Qwen2.5 VL 32B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.6,"ctx":16384,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-03-24","lu":"2025-03-24"},{"p":"meganova","id":"Qwen/Qwen3.5-Plus","n":"Qwen3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":2.4,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02","lu":"2026-02"},{"p":"meganova","id":"Qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.6,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-23","lu":"2025-07-23"},{"p":"meganova","id":"XiaomiMiMo/MiMo-V2-Flash","n":"MiMo V2 Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.1,"co":0.3,"ctx":262144,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"meganova","id":"mistralai/Mistral-Small-3.2-24B-Instruct-2506","n":"Mistral Small 3.2 24B Instruct","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0,"co":0,"ctx":32768,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-06-20","lu":"2025-06-20"},{"p":"meganova","id":"mistralai/Mistral-Nemo-Instruct-2407","n":"Mistral Nemo Instruct 2407","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.02,"co":0.04,"ctx":131072,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-18","lu":"2024-07-18"},{"p":"meganova","id":"zai-org/GLM-4.6","n":"GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.45,"co":1.9,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"meganova","id":"zai-org/GLM-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.8,"co":2.56,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"meganova","id":"zai-org/GLM-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.8,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"meganova","id":"deepseek-ai/DeepSeek-V3-0324","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.25,"co":0.88,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-03-24","lu":"2025-03-24"},{"p":"meganova","id":"deepseek-ai/DeepSeek-R1-0528","n":"DeepSeek R1 0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.15,"ctx":163840,"out":64000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-05-28","lu":"2025-05-28"},{"p":"meganova","id":"deepseek-ai/DeepSeek-V3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":164000,"out":164000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-08-25","lu":"2025-08-25"},{"p":"meganova","id":"deepseek-ai/DeepSeek-V3.2-Exp","n":"DeepSeek V3.2 Exp","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.4,"ctx":164000,"out":164000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-10-10","lu":"2025-10-10"},{"p":"meganova","id":"deepseek-ai/DeepSeek-V3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.26,"co":0.38,"ctx":164000,"out":164000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-12-03","lu":"2025-12-03"},{"p":"meganova","id":"MiniMaxAI/MiniMax-M2.1","n":"MiniMax M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.28,"co":1.2,"ctx":196608,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"meganova","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"perplexity","id":"sonar-reasoning-pro","n":"Sonar Reasoning Pro","f":"sonar-reasoning","d":"Web-grounded Sonar for multi-step research questions that need cited reasoning","ci":2,"co":8,"ctx":128000,"out":4096,"mi":"ti","mo":"t","rs":1,"kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"perplexity","id":"sonar","n":"Sonar","f":"sonar","d":"Fast web-grounded Sonar for current answers, citations, and lightweight retrieval","ci":1,"co":1,"ctx":128000,"out":4096,"mi":"t","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"perplexity","id":"sonar-pro","n":"Sonar Pro","f":"sonar-pro","d":"Deeper Sonar search model with broader retrieval and stronger synthesis","ci":3,"co":15,"ctx":200000,"out":8192,"mi":"ti","mo":"t","kn":"2025-09-01","rd":"2024-01-01","lu":"2025-09-01"},{"p":"perplexity","id":"sonar-deep-research","n":"Perplexity Sonar Deep Research","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":2,"co":8,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"kn":"2025-01","rd":"2025-02-01","lu":"2025-09-01"},{"p":"amazon-bedrock","id":"global.anthropic.claude-haiku-4-5-20251001-v1:0","n":"Claude Haiku 4.5 (Global)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"amazon-bedrock","id":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","n":"Claude Sonnet 4.5 (Global)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"amazon-bedrock","id":"jp.anthropic.claude-haiku-4-5-20251001-v1:0","n":"Claude Haiku 4.5 (JP)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"amazon-bedrock","id":"us.meta.llama4-scout-17b-instruct-v1:0","n":"Llama 4 Scout 17B Instruct (US)","f":"llama","d":"Open Llama with long-context vision for efficient multimodal agents","ci":0.17,"co":0.66,"ctx":3500000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"amazon-bedrock","id":"minimax.minimax-m2","n":"MiniMax M2","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"amazon-bedrock","id":"anthropic.claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-sonnet-4-6","n":"Claude Sonnet 4.6 (EU)","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3.3,"co":16.5,"cr":0.33,"cw":4.125,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"mistral.voxtral-small-24b-2507","n":"Voxtral Small 24B 2507","f":"mistral","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.15,"co":0.35,"ctx":32000,"out":8192,"mi":"ta","mo":"t","tc":1,"ow":1,"rd":"2025-07-01","lu":"2025-07-01"},{"p":"amazon-bedrock","id":"mistral.ministral-3-3b-instruct","n":"Ministral 3 3B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"ctx":256000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"amazon-bedrock","id":"openai.gpt-oss-20b","n":"gpt-oss-20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.07,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"amazon-bedrock","id":"anthropic.claude-opus-4-6-v1","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"openai.gpt-oss-safeguard-20b","n":"GPT OSS Safeguard 20B","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.07,"co":0.2,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-10-29","lu":"2025-10-29"},{"p":"amazon-bedrock","id":"anthropic.claude-opus-4-5-20251101-v1:0","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-08-01"},{"p":"amazon-bedrock","id":"global.anthropic.claude-fable-5","n":"Claude Fable 5 (Global)","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"amazon-bedrock","id":"openai.gpt-oss-120b-1:0","n":"gpt-oss-120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"amazon-bedrock","id":"anthropic.claude-sonnet-4-5-20250929-v1:0","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"amazon-bedrock","id":"amazon.nova-pro-v1:0","n":"Nova Pro","f":"nova-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.8,"co":3.2,"cr":0.2,"ctx":300000,"out":8192,"mi":"tiv","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-03","lu":"2024-12-03"},{"p":"amazon-bedrock","id":"qwen.qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.22,"co":1.8,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-06","lu":"2026-02-06"},{"p":"amazon-bedrock","id":"us.anthropic.claude-opus-4-7","n":"Claude Opus 4.7 (US)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"amazon-bedrock","id":"nvidia.nemotron-nano-9b-v2","n":"NVIDIA Nemotron Nano 9B v2","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0.06,"co":0.23,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"qwen.qwen3-32b-v1:0","n":"Qwen3 32B (dense)","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":0.6,"ctx":16384,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-04","rd":"2025-09-18","lu":"2025-09-18"},{"p":"amazon-bedrock","id":"jp.anthropic.claude-sonnet-4-6","n":"Claude Sonnet 4.6 (JP)","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"deepseek.r1-v1:0","n":"DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":1.35,"co":5.4,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"amazon-bedrock","id":"mistral.mistral-large-3-675b-instruct","n":"Mistral Large 3","f":"mistral","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":0.5,"co":1.5,"ctx":256000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"amazon-bedrock","id":"google.gemma-3-27b-it","n":"Google Gemma 3 27B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.12,"co":0.2,"ctx":202752,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-27","lu":"2025-07-27"},{"p":"amazon-bedrock","id":"anthropic.claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"amazon.nova-2-lite-v1:0","n":"Nova 2 Lite","f":"nova","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.33,"co":2.75,"ctx":128000,"out":4096,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"openai.gpt-oss-safeguard-120b","n":"GPT OSS Safeguard 120B","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-10-29","lu":"2025-10-29"},{"p":"amazon-bedrock","id":"mistral.ministral-3-8b-instruct","n":"Ministral 3 8B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.15,"co":0.15,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-opus-4-6-v1","n":"Claude Opus 4.6 (EU)","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5.5,"co":27.5,"cr":0.55,"cw":6.875,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"au.anthropic.claude-opus-4-6-v1","n":"AU Anthropic Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":16.5,"co":82.5,"cr":1.65,"cw":20.625,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05","rd":"2026-02-05","lu":"2026-02-05"},{"p":"amazon-bedrock","id":"jp.anthropic.claude-sonnet-5","n":"Claude Sonnet 5 (JP)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"amazon-bedrock","id":"openai.gpt-oss-120b","n":"gpt-oss-120b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"amazon-bedrock","id":"global.anthropic.claude-sonnet-5","n":"Claude Sonnet 5 (Global)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"amazon-bedrock","id":"global.anthropic.claude-opus-4-8","n":"Claude Opus 4.8 (Global)","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"amazon-bedrock","id":"us.meta.llama4-maverick-17b-instruct-v1:0","n":"Llama 4 Maverick 17B Instruct (US)","f":"llama","d":"Open multimodal Llama for strong reasoning with efficient everyday serving","ci":0.24,"co":0.97,"ctx":1000000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"amazon-bedrock","id":"openai.gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.75,"co":16.5,"cr":0.275,"ctx":272000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-06-01"},{"p":"amazon-bedrock","id":"anthropic.claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"amazon-bedrock","id":"mistral.devstral-2-123b","n":"Devstral 2 123B","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":2,"ctx":256000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-17","lu":"2026-02-17"},{"p":"amazon-bedrock","id":"zai.glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"amazon-bedrock","id":"anthropic.claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"amazon-bedrock","id":"writer.palmyra-x4-v1:0","n":"Palmyra X4","f":"palmyra","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":2.5,"co":10,"ctx":122880,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-28","lu":"2025-04-28"},{"p":"amazon-bedrock","id":"mistral.magistral-small-2509","n":"Magistral Small 1.2","f":"magistral","d":"Mistral reasoning model for transparent analysis, math, and complex decisions","ci":0.5,"co":1.5,"ctx":128000,"out":40000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"amazon-bedrock","id":"qwen.qwen3-coder-480b-a35b-v1:0","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.22,"co":1.8,"ctx":131072,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2025-09-18","lu":"2025-09-18"},{"p":"amazon-bedrock","id":"amazon.nova-micro-v1:0","n":"Nova Micro","f":"nova-micro","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.035,"co":0.14,"cr":0.00875,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-03","lu":"2024-12-03"},{"p":"amazon-bedrock","id":"mistral.pixtral-large-2502-v1:0","n":"Pixtral Large (25.02)","f":"mistral","d":"Mistral vision-language model for image understanding and multimodal chat","ci":2,"co":6,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2025-04-08","lu":"2025-04-08"},{"p":"amazon-bedrock","id":"us.anthropic.claude-opus-4-6-v1","n":"Claude Opus 4.6 (US)","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"jp.anthropic.claude-opus-4-7","n":"Claude Opus 4.7 (JP)","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"amazon-bedrock","id":"au.anthropic.claude-sonnet-4-5-20250929-v1:0","n":"Claude Sonnet 4.5 (AU)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"amazon-bedrock","id":"deepseek.v3-v1:0","n":"DeepSeek-V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.58,"co":1.68,"ctx":163840,"out":81920,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-09-18","lu":"2025-09-18"},{"p":"amazon-bedrock","id":"anthropic.claude-opus-4-1-20250805-v1:0","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"amazon-bedrock","id":"jp.anthropic.claude-sonnet-4-5-20250929-v1:0","n":"Claude Sonnet 4.5 (JP)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"amazon-bedrock","id":"google.gemma-3-4b-it","n":"Gemma 3 4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.04,"co":0.08,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-sonnet-4-5-20250929-v1:0","n":"Claude Sonnet 4.5 (EU)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3.3,"co":16.5,"cr":0.33,"cw":4.125,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"amazon-bedrock","id":"qwen.qwen3-vl-235b-a22b","n":"Qwen/Qwen3-VL-235B-A22B-Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.5,"ctx":262000,"out":262000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-04","lu":"2025-11-25"},{"p":"amazon-bedrock","id":"writer.palmyra-x5-v1:0","n":"Palmyra X5","f":"palmyra","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.6,"co":6,"ctx":1040000,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-28","lu":"2025-04-28"},{"p":"amazon-bedrock","id":"us.anthropic.claude-sonnet-4-6","n":"Claude Sonnet 4.6 (US)","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"au.anthropic.claude-haiku-4-5-20251001-v1:0","n":"Claude Haiku 4.5 (AU)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"amazon-bedrock","id":"meta.llama3-3-70b-instruct-v1:0","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.72,"co":0.72,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"amazon-bedrock","id":"zai.glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"ctx":202752,"out":101376,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"amazon-bedrock","id":"us.anthropic.claude-opus-4-8","n":"Claude Opus 4.8 (US)","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"amazon-bedrock","id":"global.anthropic.claude-opus-4-7","n":"Claude Opus 4.7 (Global)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"amazon-bedrock","id":"us.anthropic.claude-fable-5","n":"Claude Fable 5 (US)","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"amazon-bedrock","id":"amazon.nova-lite-v1:0","n":"Nova Lite","f":"nova-lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.06,"co":0.24,"cr":0.015,"ctx":300000,"out":8192,"mi":"tiv","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-03","lu":"2024-12-03"},{"p":"amazon-bedrock","id":"us.anthropic.claude-haiku-4-5-20251001-v1:0","n":"Claude Haiku 4.5 (US)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"amazon-bedrock","id":"xai.grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-06-28"},{"p":"amazon-bedrock","id":"mistral.voxtral-mini-3b-2507","n":"Voxtral Mini 3B 2507","f":"mistral","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.04,"co":0.04,"ctx":128000,"out":4096,"mi":"at","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"moonshot.kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"ctx":262143,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"amazon-bedrock","id":"meta.llama3-1-70b-instruct-v1:0","n":"Llama 3.1 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.72,"co":0.72,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"amazon-bedrock","id":"us.deepseek.r1-v1:0","n":"DeepSeek-R1 (US)","f":"deepseek-thinking","d":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","ci":1.35,"co":5.4,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"amazon-bedrock","id":"global.anthropic.claude-sonnet-4-6","n":"Claude Sonnet 4.6 (Global)","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"anthropic.claude-haiku-4-5-20251001-v1:0","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"amazon-bedrock","id":"moonshotai.kimi-k2.5","n":"Kimi K2.5","f":"kimi","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"ctx":262143,"out":16000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-06","lu":"2026-02-06"},{"p":"amazon-bedrock","id":"au.anthropic.claude-opus-4-8","n":"Claude Opus 4.8 (AU)","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"amazon-bedrock","id":"nvidia.nemotron-nano-12b-v2","n":"NVIDIA Nemotron Nano 12B v2 VL BF16","f":"nemotron","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0.2,"co":0.6,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"zai.glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"amazon-bedrock","id":"meta.llama4-scout-17b-instruct-v1:0","n":"Llama 4 Scout 17B Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.17,"co":0.66,"ctx":3500000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"amazon-bedrock","id":"au.anthropic.claude-sonnet-5","n":"Claude Sonnet 5 (AU)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"amazon-bedrock","id":"qwen.qwen3-235b-a22b-2507-v1:0","n":"Qwen3 235B A22B 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.22,"co":0.88,"ctx":262144,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2025-09-18","lu":"2025-09-18"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-haiku-4-5-20251001-v1:0","n":"Claude Haiku 4.5 (EU)","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1.1,"co":5.5,"cr":0.11,"cw":1.375,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"amazon-bedrock","id":"openai.gpt-oss-20b-1:0","n":"gpt-oss-20b","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.07,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"amazon-bedrock","id":"jp.anthropic.claude-opus-4-8","n":"Claude Opus 4.8 (JP)","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"amazon-bedrock","id":"anthropic.claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"amazon-bedrock","id":"qwen.qwen3-coder-30b-a3b-v1:0","n":"Qwen3 Coder 30B A3B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.15,"co":0.6,"ctx":262144,"out":131072,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2025-09-18","lu":"2025-09-18"},{"p":"amazon-bedrock","id":"qwen.qwen3-next-80b-a3b","n":"Qwen/Qwen3-Next-80B-A3B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.14,"co":1.4,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"rd":"2025-09-18","lu":"2025-11-25"},{"p":"amazon-bedrock","id":"openai.gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5.5,"co":33,"cr":0.55,"ctx":272000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-06-01"},{"p":"amazon-bedrock","id":"au.anthropic.claude-sonnet-4-6","n":"AU Anthropic Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3.3,"co":16.5,"cr":0.33,"cw":4.125,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08","rd":"2026-02-17","lu":"2026-02-17"},{"p":"amazon-bedrock","id":"us.anthropic.claude-sonnet-5","n":"Claude Sonnet 5 (US)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"amazon-bedrock","id":"us.anthropic.claude-opus-4-5-20251101-v1:0","n":"Claude Opus 4.5 (US)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-08-01"},{"p":"amazon-bedrock","id":"minimax.minimax-m2.5","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":196608,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-opus-4-8","n":"Claude Opus 4.8 (EU)","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5.5,"co":27.5,"cr":0.55,"cw":6.875,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-opus-4-7","n":"Claude Opus 4.7 (EU)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5.5,"co":27.5,"cr":0.55,"cw":6.875,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"amazon-bedrock","id":"meta.llama3-1-8b-instruct-v1:0","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.22,"co":0.22,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"amazon-bedrock","id":"us.anthropic.claude-opus-4-1-20250805-v1:0","n":"Claude Opus 4.1 (US)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"amazon-bedrock","id":"meta.llama4-maverick-17b-instruct-v1:0","n":"Llama 4 Maverick 17B Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.24,"co":0.97,"ctx":1000000,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"amazon-bedrock","id":"global.anthropic.claude-opus-4-5-20251101-v1:0","n":"Claude Opus 4.5 (Global)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-08-01"},{"p":"amazon-bedrock","id":"nvidia.nemotron-super-3-120b","n":"NVIDIA Nemotron 3 Super 120B A12B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.15,"co":0.65,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-sonnet-5","n":"Claude Sonnet 5 (EU)","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2.2,"co":11,"cr":0.22,"cw":2.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"amazon-bedrock","id":"nvidia.nemotron-nano-3-30b","n":"NVIDIA Nemotron Nano 3 30B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.06,"co":0.24,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-opus-4-5-20251101-v1:0","n":"Claude Opus 4.5 (EU)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5.5,"co":27.5,"cr":0.55,"cw":6.875,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-08-01"},{"p":"amazon-bedrock","id":"global.anthropic.claude-opus-4-6-v1","n":"Claude Opus 4.6 (Global)","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"amazon-bedrock","id":"us.anthropic.claude-sonnet-4-5-20250929-v1:0","n":"Claude Sonnet 4.5 (US)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"amazon-bedrock","id":"google.gemma-3-12b-it","n":"Google Gemma 3 12B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.049999999999999996,"co":0.09999999999999999,"ctx":131072,"out":8192,"mi":"ti","mo":"t","kn":"2024-12","rd":"2024-12-01","lu":"2024-12-01"},{"p":"amazon-bedrock","id":"minimax.minimax-m2.1","n":"MiniMax M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"amazon-bedrock","id":"eu.anthropic.claude-fable-5","n":"Claude Fable 5 (EU)","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":11,"co":55,"cr":1.1,"cw":13.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"amazon-bedrock","id":"deepseek.v3.2","n":"DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.62,"co":1.85,"ctx":163840,"out":81920,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2026-02-06","lu":"2026-02-06"},{"p":"amazon-bedrock","id":"mistral.ministral-3-14b-instruct","n":"Ministral 14B 3.0","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2024-12-01","lu":"2024-12-01"},{"p":"umans-ai","id":"umans-kimi-k2.7","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"umans-ai","id":"umans-glm-5.1","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0.29,"ctx":204800,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"umans-ai","id":"umans-coder","n":"Umans Coder","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"umans-ai","id":"umans-flash","n":"Umans Flash","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.15,"co":1,"cr":0.05,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"umans-ai","id":"umans-glm-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":405504,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"togetherai","id":"LiquidAI/LFM2-24B-A2B","n":"LFM2-24B-A2B","f":"liquid","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.03,"co":0.12,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2026-02-25","lu":"2026-02-25"},{"p":"togetherai","id":"meta-llama/Meta-Llama-3-8B-Instruct-Lite","n":"Meta Llama 3 8B Instruct Lite","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":0.14,"co":0.14,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2024-04-18","lu":"2024-04-18"},{"p":"togetherai","id":"meta-llama/Llama-3.3-70B-Instruct-Turbo","n":"Llama 3.3 70B","f":"llama","d":"Compact Llama instruction model for fast chat and local deployment","ci":1.04,"co":1.04,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2026-07-02"},{"p":"togetherai","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":1.2,"co":4.5,"cr":0.2,"ctx":262144,"out":131000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"togetherai","id":"moonshotai/Kimi-K2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Legacy model retained for compatibility with older integrations","ci":0.5,"co":2.8,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-01-27","lu":"2026-01-27"},{"p":"togetherai","id":"moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Kimi coding model for software agents, refactors, and repository reasoning","ci":0.95,"co":4,"cr":0.19,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-14","lu":"2026-06-14"},{"p":"togetherai","id":"google/gemma-4-31B-it","n":"Gemma 4 31B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.39,"co":0.97,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-07","lu":"2026-04-07"},{"p":"togetherai","id":"google/gemma-3n-E4B-it","n":"Gemma 3N E4B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.06,"co":0.12,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2025-05-20","lu":"2025-05-20"},{"p":"togetherai","id":"Qwen/Qwen3.7-Max","n":"Qwen3.7 Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.25,"co":3.75,"ctx":1000000,"out":500000,"mi":"t","mo":"t","tc":1,"rd":"2026-05-21","lu":"2026-07-02"},{"p":"togetherai","id":"Qwen/Qwen3.6-Plus","n":"Qwen3.6 Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.5,"co":3,"ctx":1000000,"out":500000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"togetherai","id":"Qwen/Qwen3.5-397B-A17B","n":"Qwen3.5 397B A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":262144,"out":130000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-16","lu":"2026-06-15"},{"p":"togetherai","id":"Qwen/Qwen3-Coder-Next-FP8","n":"Qwen3 Coder Next FP8","f":"qwen","d":"Legacy model retained for compatibility with older integrations","ci":0.5,"co":1.2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2026-02-03","rd":"2026-02-03","lu":"2026-02-03"},{"p":"togetherai","id":"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Legacy model retained for compatibility with older integrations","ci":2,"co":2,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"togetherai","id":"Qwen/Qwen3-235B-A22B-Instruct-2507-tput","n":"Qwen3 235B A22B Instruct 2507 FP8","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.2,"co":0.6,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-07","rd":"2025-07-25","lu":"2025-07-25"},{"p":"togetherai","id":"Qwen/Qwen2.5-7B-Instruct-Turbo","n":"Qwen 2.5 7B Instruct Turbo","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.3,"co":0.3,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-09-19","lu":"2024-09-19"},{"p":"togetherai","id":"Qwen/Qwen3.5-9B","n":"Qwen3.5 9B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.17,"co":0.25,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-03","lu":"2026-03-03"},{"p":"togetherai","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-05","lu":"2025-08-05"},{"p":"togetherai","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.2,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"togetherai","id":"pearl-ai/gemma-4-31b-it","n":"Pearl AI Gemma 4 31B Instruct","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.28,"co":0.86,"ctx":32000,"out":32000,"mi":"t","mo":"t","rs":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"togetherai","id":"nvidia/nemotron-3-ultra-550b-a55b","n":"Nemotron 3 Ultra 550B A55B","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.6,"co":3.6,"cr":0.2,"ctx":512300,"out":512300,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"togetherai","id":"deepcogito/cogito-v2-1-671b","n":"Cogito v2.1 671B","f":"cogito","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1.25,"co":1.25,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"togetherai","id":"zai-org/GLM-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":1,"co":3.2,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"togetherai","id":"zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.26,"ctx":262144,"out":164000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-16","lu":"2026-06-16"},{"p":"togetherai","id":"zai-org/GLM-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-11","rd":"2026-04-07","lu":"2026-07-02"},{"p":"togetherai","id":"deepseek-ai/DeepSeek-R1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"Legacy model retained for compatibility with older integrations","ci":3,"co":7,"ctx":163839,"out":163839,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-03-24"},{"p":"togetherai","id":"deepseek-ai/DeepSeek-V3-1","n":"DeepSeek V3.1","f":"deepseek","d":"Legacy model retained for compatibility with older integrations","ci":0.6,"co":1.7,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-21","lu":"2025-08-21"},{"p":"togetherai","id":"deepseek-ai/DeepSeek-V4-Pro","n":"DeepSeek V4 Pro","f":"deepseek","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.74,"co":3.48,"cr":0.2,"ctx":512000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"togetherai","id":"deepseek-ai/DeepSeek-V3","n":"DeepSeek-V3","f":"deepseek","d":"Legacy model retained for compatibility with older integrations","ci":1.25,"co":1.25,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-12-26","lu":"2025-05-29"},{"p":"togetherai","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":1.2,"cr":0.06,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"togetherai","id":"MiniMaxAI/MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":524288,"out":250000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-12","lu":"2026-06-12"},{"p":"togetherai","id":"MiniMaxAI/MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"togetherai","id":"essentialai/Rnj-1-Instruct","n":"Rnj-1 Instruct","f":"rnj","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.15,"co":0.15,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-12-05","lu":"2025-12-05"},{"p":"frogbot","id":"minimax-m2-5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":192000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-09","rd":"2025-01-15","lu":"2025-02-22"},{"p":"frogbot","id":"kimi-k2-6","n":"Kimi-K2.6","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":256000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"1970-01-01","lu":"1970-01-01"},{"p":"frogbot","id":"zai-glm-5-1","n":"Z.AI GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.4,"co":4.4,"cr":0.26,"ctx":198000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-01-20","lu":"2025-02-22"},{"p":"frogbot","id":"grok-code-fast-1","n":"Grok 4.1 Fast (Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":1.5,"cr":0.02,"ctx":256000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2023-10","rd":"2025-08-28","lu":"2025-08-28"},{"p":"frogbot","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.31,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-05"},{"p":"frogbot","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-07-17","lu":"2025-07-17"},{"p":"frogbot","id":"gpt-4o","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"frogbot","id":"qwen-3-6-plus","n":"Qwen 3.6 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.5,"co":3,"cr":0.1,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-02","lu":"2026-04-03"},{"p":"frogbot","id":"grok-4-3","n":"Grok 4.3","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2024-11","rd":"2026-04-30","lu":"2026-04-30"},{"p":"frogbot","id":"deepseek-v4-pro","n":"DeepSeek v4 Pro","f":"deepseek","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.74,"co":3.48,"cr":0.14,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2026-01","rd":"2026-04-24","lu":"2026-04-24"},{"p":"frogbot","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"frogbot","id":"grok-4-1-fast-non-reasoning","n":"Grok 4.1 Fast (Non-Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2025-11","rd":"2025-11-25","lu":"2025-11-25"},{"p":"frogbot","id":"minimax-m2-7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":192000,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-09","rd":"2026-03-18","lu":"2026-03-18"},{"p":"frogbot","id":"gpt-5-3-codex","n":"GPT-5.3 Codex","f":"gpt","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-02-15","lu":"2026-02-15"},{"p":"frogbot","id":"gpt-5-4-nano","n":"GPT-5.4 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"frogbot","id":"gpt-5-4-mini","n":"GPT-5.4 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"frogbot","id":"kimi-k2.5","n":"Kimi-K2.5","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":256000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"1970-01-01","lu":"1970-01-01"},{"p":"frogbot","id":"gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"1970-01-01","lu":"1970-01-01"},{"p":"frogbot","id":"gpt-5-5","n":"GPT-5.5","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"frogbot","id":"claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"frogbot","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"frogbot","id":"gemini-3-1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1000000,"out":64000,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2026-01","rd":"2026-02-18","lu":"2026-02-18"},{"p":"frogbot","id":"grok-4-1-fast-reasoning","n":"Grok 4.1 Fast (Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-11","rd":"2025-11-25","lu":"2025-11-25"},{"p":"frogbot","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"frogbot","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"frogbot","id":"gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.07,"co":0.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"1970-01-01","lu":"1970-01-01"},{"p":"openrouter","id":"inclusionai/ling-2.6-1t","n":"Ling-2.6-1T","f":"ling","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.075,"co":0.625,"cr":0.015,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-04-23","lu":"2026-04-23"},{"p":"openrouter","id":"inclusionai/ring-2.6-1t","n":"Ring-2.6-1T","f":"ring","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.075,"co":0.625,"cr":0.015,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-08","lu":"2026-05-08"},{"p":"openrouter","id":"inclusionai/ling-2.6-flash","n":"Ling-2.6-flash","f":"ling","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.01,"co":0.03,"cr":0.002,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"openrouter","id":"ibm-granite/granite-4.0-h-micro","n":"Granite 4.0 Micro","f":"granite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.017,"co":0.112,"ctx":131000,"out":131000,"mi":"t","mo":"t","ow":1,"rd":"2025-10-20","lu":"2025-10-20"},{"p":"openrouter","id":"ibm-granite/granite-4.1-8b","n":"Granite 4.1 8B","f":"granite","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.05,"co":0.1,"cr":0.05,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"openrouter","id":"meta-llama/llama-3.1-8b-instruct","n":"Llama 3.1 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.02,"co":0.03,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12-31","rd":"2024-07-23","lu":"2024-07-23"},{"p":"openrouter","id":"meta-llama/llama-3.1-70b-instruct","n":"Llama 3.1 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.4,"co":0.4,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12-31","rd":"2024-07-23","lu":"2024-07-23"},{"p":"openrouter","id":"meta-llama/llama-3.2-1b-instruct","n":"Llama 3.2 1B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.027,"co":0.201,"ctx":60000,"out":60000,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-09-25","lu":"2024-09-25"},{"p":"openrouter","id":"meta-llama/llama-4-maverick","n":"Llama 4 Maverick","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.15,"co":0.6,"ctx":1048576,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08-31","rd":"2025-04-05","lu":"2025-04-05"},{"p":"openrouter","id":"meta-llama/llama-3.2-11b-vision-instruct","n":"Llama 3.2 11B Vision Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.345,"co":0.345,"ctx":131072,"out":16384,"mi":"ti","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-09-25","lu":"2024-09-25"},{"p":"openrouter","id":"meta-llama/llama-3.3-70b-instruct:free","n":"Llama 3.3 70B Instruct (free)","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":0,"co":0,"ctx":65536,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"openrouter","id":"meta-llama/llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Popular open Llama workhorse for multilingual chat, coding, and self-hosting","ci":0.1,"co":0.32,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"openrouter","id":"meta-llama/llama-3.2-3b-instruct:free","n":"Llama 3.2 3B Instruct (free)","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-09-25","lu":"2024-09-25"},{"p":"openrouter","id":"meta-llama/llama-guard-4-12b","n":"Llama Guard 4 12B","f":"llama","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.18,"co":0.18,"ctx":163840,"out":16384,"mi":"it","mo":"t","ow":1,"kn":"2024-08-31","rd":"2025-04-30","lu":"2025-04-30"},{"p":"openrouter","id":"meta-llama/llama-3-8b-instruct","n":"Llama 3 8B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.14,"co":0.14,"ctx":8192,"out":8192,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-04-18","lu":"2024-04-18"},{"p":"openrouter","id":"meta-llama/llama-4-scout","n":"Llama 4 Scout","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.1,"co":0.3,"ctx":327680,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08-31","rd":"2025-04-05","lu":"2025-04-05"},{"p":"openrouter","id":"meta-llama/llama-3.2-3b-instruct","n":"Llama 3.2 3B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.05,"co":0.33,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-09-25","lu":"2024-09-25"},{"p":"openrouter","id":"~anthropic/claude-haiku-latest","n":"Anthropic Claude Haiku Latest","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"~anthropic/claude-fable-latest","n":"Claude Fable Latest","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-06-09","lu":"2026-06-09"},{"p":"openrouter","id":"~anthropic/claude-sonnet-latest","n":"Anthropic Claude Sonnet Latest","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"~anthropic/claude-opus-latest","n":"Claude Opus Latest","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"openrouter","id":"moonshotai/kimi-k2","n":"Kimi K2 0711","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.57,"co":2.3,"ctx":131072,"out":100352,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12-31","rd":"2025-07-11","lu":"2025-07-11"},{"p":"openrouter","id":"moonshotai/kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.74,"co":3.5,"cr":0.15,"ctx":262144,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"openrouter","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Thinking Kimi model for slower research passes, planning, and hard technical questions","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":100352,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-11-06"},{"p":"openrouter","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.375,"co":2.025,"cr":0.203,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"openrouter","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.66,"co":3.41,"cr":0.14,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"openrouter","id":"moonshotai/kimi-k2-0905","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"ctx":262144,"out":100352,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12-31","rd":"2025-09-04","lu":"2025-09-04"},{"p":"openrouter","id":"baidu/ernie-4.5-vl-424b-a47b","n":"ERNIE 4.5 VL 424B A47B ","f":"ernie","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.42,"co":1.25,"ctx":123000,"out":16000,"mi":"it","mo":"t","rs":1,"ow":1,"kn":"2025-03-31","rd":"2025-06-30","lu":"2025-06-30"},{"p":"openrouter","id":"perceptron/perceptron-mk1","n":"Perceptron Mk1","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.15,"co":1.5,"ctx":32768,"out":8192,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"openrouter","id":"google/gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":0.083333,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"openrouter","id":"google/gemini-3.1-flash-image","n":"Nano Banana 2 (Gemini 3.1 Flash Image)","f":"gemini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":3,"ctx":131072,"out":32768,"mi":"it","mo":"it","rs":1,"rd":"2026-06-18","lu":"2026-06-18"},{"p":"openrouter","id":"google/gemma-3n-e4b-it","n":"Gemma 3n 4B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.06,"co":0.12,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"kn":"2024-08-31","rd":"2025-05-20","lu":"2025-05-20"},{"p":"openrouter","id":"google/gemma-4-26b-a4b-it:free","n":"Gemma 4 26B A4B  (free)","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"openrouter","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"cw":0.375,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"openrouter","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"cw":0.083333,"ctx":1048576,"out":65535,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"openrouter","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"cw":0.083333,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"openrouter","id":"google/gemini-2.5-flash-lite-preview-09-2025","n":"Gemini 2.5 Flash Lite Preview 09-2025","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.01,"cw":0.083333,"ctx":1048576,"out":65535,"mi":"tipav","mo":"t","rs":1,"tc":1,"kn":"2025-01-31","rd":"2025-09-25","lu":"2025-09-25"},{"p":"openrouter","id":"google/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.12,"co":0.35,"cr":0.09,"ctx":262144,"out":262144,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"openrouter","id":"google/lyria-3-clip-preview","n":"Lyria 3 Clip Preview","f":"lyria","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":1048576,"out":65536,"mi":"ti","mo":"ta","rd":"2026-03-30","lu":"2026-03-30"},{"p":"openrouter","id":"google/gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"cw":0.375,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"openrouter","id":"google/gemini-3.1-flash-lite-image","n":"Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)","f":"gemini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.25,"co":1.5,"ctx":65536,"out":66000,"mi":"it","mo":"it","rs":1,"kn":"2025-01-01","rd":"2026-06-30","lu":"2026-06-30"},{"p":"openrouter","id":"google/gemini-3-pro-image-preview","n":"Nano Banana Pro","f":"gemini-pro","d":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","ci":2,"co":12,"cr":0.2,"cw":0.375,"ctx":65536,"out":32768,"mi":"ti","mo":"ti","rs":1,"kn":"2025-01","rd":"2025-11-20","lu":"2025-11-20"},{"p":"openrouter","id":"google/gemini-2.5-flash-image","n":"Nano Banana","f":"gemini-flash","d":"Nano Banana image model for fast generation, edits, and character-consistent assets","ci":0.3,"co":2.5,"cr":0.03,"cw":0.083333,"ctx":32768,"out":32768,"mi":"ti","mo":"ti","kn":"2025-06","rd":"2025-08-26","lu":"2025-08-26"},{"p":"openrouter","id":"google/gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"cw":0.083333,"ctx":1048576,"out":65535,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"openrouter","id":"google/gemini-3.1-flash-image-preview","n":"Nano Banana 2","f":"gemini-flash","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":3,"ctx":131072,"out":32768,"mi":"it","mo":"ti","rs":1,"kn":"2025-01","rd":"2026-02-26","lu":"2026-02-26"},{"p":"openrouter","id":"google/gemini-2.5-pro-preview-05-06","n":"Gemini 2.5 Pro Preview 05-06","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"cw":0.375,"ctx":1048576,"out":65535,"mi":"tipav","mo":"t","rs":1,"tc":1,"kn":"2025-01-31","rd":"2025-05-07","lu":"2025-05-07"},{"p":"openrouter","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"cw":0.375,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"openrouter","id":"google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B IT","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.06,"co":0.33,"ctx":262144,"out":262144,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"openrouter","id":"google/gemini-2.5-pro-preview","n":"Gemini 2.5 Pro Preview 06-05","f":"gemini","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"cw":0.375,"ctx":1048576,"out":65536,"mi":"pita","mo":"t","rs":1,"tc":1,"kn":"2025-01-31","rd":"2025-06-05","lu":"2025-06-05"},{"p":"openrouter","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"cw":0.083333,"ctx":1048576,"out":65535,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"openrouter","id":"google/gemma-3-12b-it","n":"Gemma 3 12B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.05,"co":0.15,"ctx":131072,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08-31","rd":"2025-03-13","lu":"2025-03-13"},{"p":"openrouter","id":"google/gemma-3-4b-it","n":"Gemma 3 4B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.05,"co":0.1,"ctx":131072,"out":16384,"mi":"ti","mo":"t","ow":1,"kn":"2024-08-31","rd":"2025-03-13","lu":"2025-03-13"},{"p":"openrouter","id":"google/gemma-3-27b-it","n":"Gemma 3 27B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.08,"co":0.16,"ctx":131072,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08-31","rd":"2025-03-12","lu":"2025-03-12"},{"p":"openrouter","id":"google/lyria-3-pro-preview","n":"Lyria 3 Pro Preview","f":"lyria","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":1048576,"out":65536,"mi":"ti","mo":"ta","rd":"2026-03-30","lu":"2026-03-30"},{"p":"openrouter","id":"google/gemma-4-31b-it:free","n":"Gemma 4 31B (free)","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0,"co":0,"ctx":262144,"out":8192,"mi":"itv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"openrouter","id":"google/gemma-2-27b-it","n":"Gemma 2 27B","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.65,"co":0.65,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2024-06-30","rd":"2024-07-13","lu":"2024-07-13"},{"p":"openrouter","id":"google/gemini-3-pro-image","n":"Nano Banana Pro (Gemini 3 Pro Image)","f":"gemini","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":12,"cr":0.2,"cw":0.375,"ctx":65536,"out":32768,"mi":"it","mo":"it","rs":1,"tc":1,"rd":"2026-06-18","lu":"2026-06-18"},{"p":"openrouter","id":"google/gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":0.083333,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"openrouter","id":"liquid/lfm-2.5-1.2b-thinking:free","n":"LFM2.5-1.2B-Thinking (free)","f":"liquid","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":32768,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-01-20","lu":"2026-01-20"},{"p":"openrouter","id":"liquid/lfm-2-24b-a2b","n":"LFM2-24B-A2B","f":"liquid","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.03,"co":0.12,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"rd":"2026-02-25","lu":"2026-02-25"},{"p":"openrouter","id":"liquid/lfm-2.5-1.2b-instruct:free","n":"LFM2.5-1.2B-Instruct (free)","f":"liquid","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"kn":"2025-06","rd":"2026-01-20","lu":"2026-01-20"},{"p":"openrouter","id":"x-ai/grok-4.20","n":"Grok 4.20","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-09-01","rd":"2026-03-31","lu":"2026-03-31"},{"p":"openrouter","id":"x-ai/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":1000000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"openrouter","id":"x-ai/grok-4.20-multi-agent","n":"Grok 4.20 Multi-Agent","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":2000000,"out":2000000,"mi":"tip","mo":"t","rs":1,"kn":"2025-09-01","rd":"2026-03-31","lu":"2026-03-31"},{"p":"openrouter","id":"x-ai/grok-build-0.1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"openrouter","id":"~google/gemini-pro-latest","n":"Google Gemini Pro Latest","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"cw":0.375,"ctx":1048576,"out":65536,"mi":"apitv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"~google/gemini-flash-latest","n":"Google Gemini Flash Latest","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"cw":0.083333,"ctx":1048576,"out":65536,"mi":"tivpa","mo":"t","rs":1,"tc":1,"kn":"2025-01-01","rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"microsoft/phi-4","n":"Phi 4","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.07,"co":0.14,"ctx":16384,"out":16384,"mi":"t","mo":"t","ow":1,"kn":"2024-06-30","rd":"2025-01-10","lu":"2025-01-10"},{"p":"openrouter","id":"microsoft/wizardlm-2-8x22b","n":"WizardLM-2 8x22B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.62,"co":0.62,"ctx":65535,"out":8000,"mi":"t","mo":"t","ow":1,"kn":"2024-04-30","rd":"2024-04-16","lu":"2024-04-16"},{"p":"openrouter","id":"poolside/laguna-xs-2.1:free","n":"Laguna XS 2.1 (free)","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-07-02","lu":"2026-07-02"},{"p":"openrouter","id":"poolside/laguna-xs.2","n":"Laguna XS.2","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.1,"co":0.2,"cr":0.05,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"openrouter","id":"poolside/laguna-xs.2:free","n":"Laguna XS.2 (free)","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"openrouter","id":"poolside/laguna-m.1","n":"Laguna M.1","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.2,"co":0.4,"cr":0.1,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"openrouter","id":"poolside/laguna-xs-2.1","n":"Laguna XS 2.1","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.06,"co":0.12,"cr":0.03,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-07-02","lu":"2026-07-02"},{"p":"openrouter","id":"poolside/laguna-m.1:free","n":"Laguna M.1 (free)","d":"Free provider route for experiments, demos, and cost-sensitive chat workloads","ci":0,"co":0,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"openrouter","id":"writer/palmyra-x5","n":"Palmyra X5","f":"palmyra","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.6,"co":6,"ctx":1040000,"out":8192,"mi":"t","mo":"t","rd":"2026-01-21","lu":"2026-01-21"},{"p":"openrouter","id":"z-ai/glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.4,"co":1.75,"cr":0.08,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"openrouter","id":"z-ai/glm-4.5v","n":"GLM-4.5V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"cr":0.11,"ctx":65536,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-08-11","lu":"2025-08-11"},{"p":"openrouter","id":"z-ai/glm-4.5","n":"GLM-4.5","f":"glm","d":"Hybrid-reasoning GLM release that made the 4.5 line broadly useful","ci":0.6,"co":2.2,"cr":0.11,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"openrouter","id":"z-ai/glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.966,"co":3.036,"cr":0.1794,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"openrouter","id":"z-ai/glm-4.6","n":"GLM-4.6","f":"glm","d":"Late GLM-4 workhorse for coding agents, reasoning, and structured tasks","ci":0.43,"co":1.74,"cr":0.08,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09-30","lu":"2025-09-30"},{"p":"openrouter","id":"z-ai/glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0.9086,"co":2.8556,"cr":0.16874,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"openrouter","id":"z-ai/glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"cr":0.055,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"openrouter","id":"z-ai/glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"Fast GLM vision model for screenshots, documents, and multimodal agent tasks","ci":1.2,"co":4,"cr":0.24,"ctx":202752,"out":131072,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"openrouter","id":"z-ai/glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Lighter GLM-4.5 variant for fast coding assistance and cheaper agents","ci":0.13,"co":0.85,"cr":0.025,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"openrouter","id":"z-ai/glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0.06,"co":0.4,"cr":0.01,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"openrouter","id":"z-ai/glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.6,"co":1.92,"cr":0.12,"ctx":202752,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"openrouter","id":"z-ai/glm-5-turbo","n":"GLM-5-Turbo","f":"glm","d":"Faster GLM-5 lane for coding agents that need lower latency","ci":1.2,"co":4,"cr":0.24,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"openrouter","id":"openai/gpt-4o-mini-2024-07-18","n":"GPT-4o-mini (2024-07-18)","f":"o-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-10-31","rd":"2024-07-18","lu":"2024-07-18"},{"p":"openrouter","id":"openai/gpt-oss-safeguard-20b","n":"gpt-oss-safeguard-20b","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.075,"co":0.3,"cr":0.0375,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-29","lu":"2025-10-29"},{"p":"openrouter","id":"openai/gpt-3.5-turbo-instruct","n":"GPT-3.5 Turbo Instruct","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":4095,"out":4096,"mi":"t","mo":"t","kn":"2021-09-30","rd":"2023-09-28","lu":"2023-09-28"},{"p":"openrouter","id":"openai/gpt-5.2-chat","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"pit","mo":"t","tc":1,"kn":"2025-08-31","rd":"2025-12-10","lu":"2025-12-10"},{"p":"openrouter","id":"openai/o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"openrouter","id":"openai/o4-mini-high","n":"o4 Mini High","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2024-06-30","rd":"2025-04-16","lu":"2025-04-16"},{"p":"openrouter","id":"openai/gpt-audio","n":"GPT Audio","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"ta","mo":"ta","tc":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"openrouter","id":"openai/gpt-5.2-pro","n":"GPT-5.2 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5.2 variant for tougher reasoning and review workflows","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openrouter","id":"openai/gpt-4o-mini-search-preview","n":"GPT-4o-mini Search Preview","f":"o-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":128000,"out":16384,"mi":"t","mo":"t","kn":"2023-10-31","rd":"2025-03-12","lu":"2025-03-12"},{"p":"openrouter","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openrouter","id":"openai/gpt-5-chat","n":"GPT-5 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"pit","mo":"t","kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openrouter","id":"openai/gpt-3.5-turbo","n":"GPT-3.5-turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"ctx":16385,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2021-09-01","rd":"2023-03-01","lu":"2023-11-06"},{"p":"openrouter","id":"openai/gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":128000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"openrouter","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"openrouter","id":"openai/gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":30,"co":60,"ctx":8191,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"openrouter","id":"openai/o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"openrouter","id":"openai/gpt-3.5-turbo-16k","n":"GPT-3.5 Turbo 16k","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":4,"ctx":16385,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2021-09-30","rd":"2023-08-28","lu":"2023-08-28"},{"p":"openrouter","id":"openai/o3-pro","n":"o3-pro","f":"o-pro","d":"High-effort o3 tier for difficult technical reasoning and careful answers","ci":20,"co":80,"ctx":200000,"out":100000,"mi":"tpi","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-06-10","lu":"2025-06-10"},{"p":"openrouter","id":"openai/gpt-5.1-chat","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.13,"ctx":128000,"out":32000,"mi":"pit","mo":"t","tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openrouter","id":"openai/gpt-4o-2024-05-13","n":"GPT-4o (2024-05-13)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":15,"ctx":128000,"out":4096,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-05-13"},{"p":"openrouter","id":"openai/gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"openrouter","id":"openai/gpt-5.3-chat","n":"GPT-5.3 Chat","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"rd":"2026-03-03","lu":"2026-03-03"},{"p":"openrouter","id":"openai/gpt-3.5-turbo-0613","n":"GPT-3.5 Turbo (older v0613)","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":2,"ctx":4095,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2021-09-30","rd":"2024-01-25","lu":"2024-01-25"},{"p":"openrouter","id":"openai/gpt-5-image-mini","n":"GPT-5 Image Mini","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2.5,"co":2,"cr":0.25,"ctx":400000,"out":128000,"mi":"pit","mo":"it","rs":1,"rd":"2025-10-16","lu":"2025-10-16"},{"p":"openrouter","id":"openai/gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Codex GPT for repository edits, code review, and practical software agents","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openrouter","id":"openai/gpt-5.1-codex-max","n":"GPT-5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openrouter","id":"openai/gpt-oss-120b:free","n":"gpt-oss-120b (free)","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0,"co":0,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"openrouter","id":"openai/gpt-4o-2024-08-06","n":"GPT-4o (2024-08-06)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-08-06","lu":"2024-08-06"},{"p":"openrouter","id":"openai/o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"tp","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"openrouter","id":"openai/gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openrouter","id":"openai/gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"openrouter","id":"openai/gpt-audio-mini","n":"GPT Audio Mini","f":"o-mini","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0.6,"co":2.4,"ctx":128000,"out":16384,"mi":"ta","mo":"ta","tc":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"openrouter","id":"openai/gpt-5.1-codex-mini","n":"GPT-5.1 Codex mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openrouter","id":"openai/o4-mini-deep-research","n":"o4-mini-deep-research","f":"o-mini","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-06-26","lu":"2024-06-26"},{"p":"openrouter","id":"openai/gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"itp","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"openrouter","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.03,"co":0.15,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"openrouter","id":"openai/gpt-4o-2024-11-20","n":"GPT-4o (2024-11-20)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-11-20","lu":"2024-11-20"},{"p":"openrouter","id":"openai/o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"openrouter","id":"openai/o1-pro","n":"o1-pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":150,"co":600,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rs":1,"kn":"2023-09","rd":"2025-03-19","lu":"2025-03-19"},{"p":"openrouter","id":"openai/gpt-chat-latest","n":"GPT Chat Latest","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":5,"co":30,"cr":0.5,"ctx":400000,"out":128000,"mi":"tip","mo":"t","tc":1,"rd":"2026-05-05","lu":"2026-05-05"},{"p":"openrouter","id":"openai/gpt-5-image","n":"GPT-5 Image","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":10,"co":10,"cr":1.25,"ctx":400000,"out":128000,"mi":"itp","mo":"it","rs":1,"kn":"2024-10-01","rd":"2025-10-14","lu":"2025-10-14"},{"p":"openrouter","id":"openai/gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"openrouter","id":"openai/gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"openrouter","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"openrouter","id":"openai/o3-deep-research","n":"o3-deep-research","f":"o","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":10,"co":40,"cr":2.5,"ctx":200000,"out":100000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-06-26","lu":"2024-06-26"},{"p":"openrouter","id":"openai/gpt-4-turbo-preview","n":"GPT-4 Turbo Preview","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2023-12-31","rd":"2024-01-25","lu":"2024-01-25"},{"p":"openrouter","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openrouter","id":"openai/gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"openrouter","id":"openai/gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"openrouter","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.01,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"openrouter","id":"openai/gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"openrouter","id":"openai/o3-mini-high","n":"o3 Mini High","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"tp","mo":"t","rs":1,"tc":1,"kn":"2023-10-31","rd":"2025-02-12","lu":"2025-02-12"},{"p":"openrouter","id":"openai/gpt-5.4-image-2","n":"GPT-5.4 Image 2","f":"gpt","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":8,"co":15,"cr":2,"ctx":272000,"out":128000,"mi":"itp","mo":"it","rs":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"openrouter","id":"openai/gpt-4o-search-preview","n":"GPT-4o Search Preview","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"t","mo":"t","kn":"2023-10-31","rd":"2025-03-12","lu":"2025-03-12"},{"p":"openrouter","id":"openai/gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"openrouter","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"openrouter","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.029,"co":0.14,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"openrouter","id":"openai/gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"openrouter","id":"openai/gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"openrouter","id":"openai/gpt-oss-20b:free","n":"gpt-oss-20b (free)","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0,"co":0,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"openrouter","id":"openai/gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"openrouter","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"openrouter","id":"thedrummer/cydonia-24b-v4.1","n":"Cydonia 24B V4.1","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.3,"co":0.5,"cr":0.15,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"kn":"2024-04-30","rd":"2025-09-27","lu":"2025-09-27"},{"p":"openrouter","id":"thedrummer/skyfall-36b-v2","n":"Skyfall 36B V2","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.55,"co":0.8,"cr":0.25,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"kn":"2024-06-30","rd":"2025-03-10","lu":"2025-03-10"},{"p":"openrouter","id":"thedrummer/unslopnemo-12b","n":"UnslopNemo 12B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.4,"co":0.4,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04-30","rd":"2024-11-08","lu":"2024-11-08"},{"p":"openrouter","id":"thedrummer/rocinante-12b","n":"Rocinante 12B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.25,"co":0.5,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"kn":"2024-04-30","rd":"2024-09-30","lu":"2024-09-30"},{"p":"openrouter","id":"bytedance/ui-tars-1.5-7b","n":"UI-TARS 7B ","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.1,"co":0.2,"cr":0.1,"ctx":128000,"out":2048,"mi":"it","mo":"t","ow":1,"kn":"2025-01-31","rd":"2025-07-22","lu":"2025-07-22"},{"p":"openrouter","id":"rekaai/reka-flash-3","n":"Reka Flash 3","f":"reka","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.1,"co":0.2,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2025-01-31","rd":"2025-03-12","lu":"2025-03-12"},{"p":"openrouter","id":"rekaai/reka-edge","n":"Reka Edge","f":"reka","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.1,"co":0.1,"ctx":16384,"out":16384,"mi":"itv","mo":"t","tc":1,"ow":1,"rd":"2026-03-20","lu":"2026-03-20"},{"p":"openrouter","id":"mistralai/mistral-large-2407","n":"Mistral Large 2407","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"cr":0.2,"ctx":131072,"out":131072,"mi":"tp","mo":"t","tc":1,"kn":"2024-03-31","rd":"2024-11-19","lu":"2024-11-19"},{"p":"openrouter","id":"mistralai/mistral-small-3.2-24b-instruct","n":"Mistral Small 3.2 24B","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.075,"co":0.2,"ctx":128000,"out":16384,"mi":"it","mo":"t","tc":1,"ow":1,"kn":"2023-10-31","rd":"2025-06-20","lu":"2025-06-20"},{"p":"openrouter","id":"mistralai/mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","ci":0.02,"co":0.03,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-07-01","lu":"2024-07-01"},{"p":"openrouter","id":"mistralai/mistral-medium-3-5","n":"Mistral Medium 3.5","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":1.5,"co":7.5,"ctx":262144,"out":262144,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"openrouter","id":"mistralai/ministral-8b-2512","n":"Ministral 3 8B 2512","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.15,"co":0.15,"cr":0.015,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"openrouter","id":"mistralai/mistral-small-3.1-24b-instruct","n":"Mistral Small 3.1 24B","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.351,"co":0.555,"ctx":128000,"out":128000,"mi":"ti","mo":"t","ow":1,"kn":"2023-10-31","rd":"2025-03-17","lu":"2025-03-17"},{"p":"openrouter","id":"mistralai/mistral-saba","n":"Saba","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.2,"co":0.6,"cr":0.02,"ctx":32768,"out":32768,"mi":"tp","mo":"t","tc":1,"kn":"2024-09-30","rd":"2025-02-17","lu":"2025-02-17"},{"p":"openrouter","id":"mistralai/mistral-large","n":"Mistral Large","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"cr":0.2,"ctx":128000,"out":128000,"mi":"tp","mo":"t","tc":1,"kn":"2024-11-30","rd":"2024-02-26","lu":"2024-02-26"},{"p":"openrouter","id":"mistralai/mistral-medium-3.1","n":"Mistral Medium 3.1","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"cr":0.04,"ctx":131072,"out":262144,"mi":"tip","mo":"t","tc":1,"kn":"2025-06-30","rd":"2025-08-13","lu":"2025-08-13"},{"p":"openrouter","id":"mistralai/mistral-small-24b-instruct-2501","n":"Mistral Small 3","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.05,"co":0.08,"ctx":32768,"out":16384,"mi":"t","mo":"t","ow":1,"kn":"2023-10-31","rd":"2025-01-30","lu":"2025-01-30"},{"p":"openrouter","id":"mistralai/ministral-3b-2512","n":"Ministral 3 3B 2512","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"cr":0.01,"ctx":131072,"out":131072,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"openrouter","id":"mistralai/mistral-small-2603","n":"Mistral Small 4","f":"mistral-small","d":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","ci":0.15,"co":0.6,"cr":0.015,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06","rd":"2026-03-16","lu":"2026-03-16"},{"p":"openrouter","id":"mistralai/ministral-14b-2512","n":"Ministral 3 14B 2512","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"cr":0.02,"ctx":262144,"out":262144,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"openrouter","id":"mistralai/devstral-2512","n":"Devstral 2","f":"devstral","d":"Mistral's coding-agent model for repository work, terminal tasks, and software fixes","ci":0.4,"co":2,"cr":0.04,"ctx":262144,"out":262144,"mi":"tp","mo":"t","tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-09","lu":"2025-12-09"},{"p":"openrouter","id":"mistralai/mixtral-8x22b-instruct","n":"Mixtral 8x22B Instruct","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":2,"co":6,"cr":0.2,"ctx":65536,"out":65536,"mi":"tp","mo":"t","tc":1,"ow":1,"kn":"2024-01-31","rd":"2024-04-17","lu":"2024-04-17"},{"p":"openrouter","id":"mistralai/mistral-medium-3","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"cr":0.04,"ctx":131072,"out":131072,"mi":"tip","mo":"t","tc":1,"kn":"2025-03-31","rd":"2025-05-07","lu":"2025-05-07"},{"p":"openrouter","id":"mistralai/voxtral-small-24b-2507","n":"Voxtral Small 24B 2507","f":"mistral","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"cr":0.01,"ctx":32000,"out":32000,"mi":"tap","mo":"t","tc":1,"ow":1,"rd":"2025-10-30","lu":"2025-10-30"},{"p":"openrouter","id":"mistralai/mistral-large-2512","n":"Mistral Large 3","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ci":0.5,"co":1.5,"cr":0.05,"ctx":262144,"out":262144,"mi":"tip","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"openrouter","id":"mistralai/codestral-2508","n":"Codestral 2508","f":"codestral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.9,"cr":0.03,"ctx":256000,"out":256000,"mi":"tp","mo":"t","tc":1,"kn":"2025-03-31","rd":"2025-08-01","lu":"2025-08-01"},{"p":"openrouter","id":"morph/morph-v3-fast","n":"Morph V3 Fast","f":"morph","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.8,"co":1.2,"ctx":81920,"out":38000,"mi":"t","mo":"t","rd":"2025-07-07","lu":"2025-07-07"},{"p":"openrouter","id":"morph/morph-v3-large","n":"Morph V3 Large","f":"morph","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.9,"co":1.9,"ctx":262144,"out":131072,"mi":"t","mo":"t","rd":"2025-07-07","lu":"2025-07-07"},{"p":"openrouter","id":"bytedance-seed/seed-1.6-flash","n":"Seed 1.6 Flash","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.075,"co":0.3,"ctx":262144,"out":32768,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"openrouter","id":"bytedance-seed/seed-1.6","n":"Seed 1.6","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"ctx":262144,"out":32768,"mi":"itv","mo":"t","rs":1,"tc":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"openrouter","id":"bytedance-seed/seed-2.0-mini","n":"Seed-2.0-Mini","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.1,"co":0.4,"ctx":262144,"out":131072,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-26","lu":"2026-02-26"},{"p":"openrouter","id":"bytedance-seed/seed-2.0-lite","n":"Seed-2.0-Lite","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":2,"ctx":262144,"out":131072,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-03-10","lu":"2026-03-10"},{"p":"openrouter","id":"sakana/fugu-ultra","n":"Fugu Ultra","f":"fugu","d":"Quality-first multi-agent model for hard research, analysis, and competitions","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-24","lu":"2026-06-24"},{"p":"openrouter","id":"anthracite-org/magnum-v4-72b","n":"Magnum v4 72B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":3,"co":5,"ctx":16384,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2024-06-30","rd":"2024-10-22","lu":"2024-10-22"},{"p":"openrouter","id":"nvidia/nemotron-3-nano-30b-a3b:free","n":"Nemotron 3 Nano 30B A3B (free)","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0,"co":0,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-15","lu":"2025-12-15"},{"p":"openrouter","id":"nvidia/nemotron-nano-9b-v2:free","n":"Nemotron Nano 9B V2 (free)","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0,"co":0,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-18","lu":"2025-08-18"},{"p":"openrouter","id":"nvidia/nemotron-nano-12b-v2-vl:free","n":"Nemotron Nano 12B 2 VL (free)","f":"nemotron","d":"Nemotron multimodal model for visual reasoning and agentic AI workflows","ci":0,"co":0,"ctx":128000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-28","lu":"2025-10-28"},{"p":"openrouter","id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","n":"Nemotron 3 Nano Omni (free)","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0,"co":0,"ctx":256000,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"openrouter","id":"nvidia/nemotron-3-ultra-550b-a55b:free","n":"Nemotron 3 Ultra (free)","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0,"co":0,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"openrouter","id":"nvidia/nemotron-3-ultra-550b-a55b","n":"Nemotron 3 Ultra 550B A55B","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.5,"co":2.2,"cr":0.1,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"openrouter","id":"nvidia/nemotron-3.5-content-safety:free","n":"Nemotron 3.5 Content Safety (free)","f":"nemotron","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0,"co":0,"ctx":128000,"out":8192,"mi":"ti","mo":"t","rs":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"openrouter","id":"nvidia/nemotron-3-nano-30b-a3b","n":"Nemotron 3 Nano 30B A3B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.05,"co":0.2,"ctx":262144,"out":228000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-15","lu":"2025-12-15"},{"p":"openrouter","id":"nvidia/llama-3.3-nemotron-super-49b-v1.5","n":"Llama 3.3 Nemotron Super 49B v1.5","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.4,"co":0.4,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-25","lu":"2025-07-25"},{"p":"openrouter","id":"nvidia/nemotron-3-super-120b-a12b:free","n":"Nemotron 3 Super (free)","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"openrouter","id":"nvidia/nemotron-3-super-120b-a12b","n":"Nemotron 3 Super 120B A12B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.08,"co":0.45,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"openrouter","id":"cognitivecomputations/dolphin-mistral-24b-venice-edition:free","n":"Uncensored (free)","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0,"co":0,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"kn":"2024-04-30","rd":"2025-07-09","lu":"2025-07-09"},{"p":"openrouter","id":"xiaomi/mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.105,"co":0.28,"cr":0.028,"ctx":32000,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"openrouter","id":"xiaomi/mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"openrouter","id":"inception/mercury-2","n":"Mercury 2","f":"mercury","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.25,"co":0.75,"cr":0.025,"ctx":128000,"out":50000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-04","lu":"2026-03-04"},{"p":"openrouter","id":"anthropic/claude-sonnet-4.5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"openrouter","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"openrouter","id":"anthropic/claude-haiku-4.5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"openrouter","id":"anthropic/claude-opus-4.7-fast","n":"Claude Opus 4.7 (Fast)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":30,"co":150,"cr":3,"cw":37.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"openrouter","id":"anthropic/claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"openrouter","id":"anthropic/claude-opus-4.7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"openrouter","id":"anthropic/claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"openrouter","id":"anthropic/claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"openrouter","id":"anthropic/claude-opus-4.1","n":"Claude Opus 4.1 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"openrouter","id":"anthropic/claude-opus-4.5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"openrouter","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"openrouter","id":"anthropic/claude-3-haiku","n":"Claude 3 Haiku","f":"claude","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.25,"co":1.25,"cr":0.03,"cw":0.3,"ctx":200000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-08-31","rd":"2024-03-13","lu":"2024-03-13"},{"p":"openrouter","id":"anthropic/claude-opus-4","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"itp","mo":"t","rs":1,"tc":1,"kn":"2025-01-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"openrouter","id":"anthropic/claude-opus-4.8-fast","n":"Claude Opus 4.8 (Fast)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-27","lu":"2026-05-27"},{"p":"openrouter","id":"anthropic/claude-opus-4.6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"openrouter","id":"tencent/hy3:free","n":"Hy3 (free)","f":"hy3","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-07-06","lu":"2026-07-06"},{"p":"openrouter","id":"tencent/hy3","n":"Hy3","f":"hy3","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.14,"co":0.58,"cr":0.035,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-07-06","lu":"2026-07-06"},{"p":"openrouter","id":"tencent/hunyuan-a13b-instruct","n":"Hunyuan A13B Instruct","f":"hunyuan","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.14,"co":0.57,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2025-03-31","rd":"2025-07-08","lu":"2025-07-08"},{"p":"openrouter","id":"tencent/hy3-preview","n":"Hy3 preview","f":"Hy","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.063,"co":0.21,"cr":0.021,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"openrouter","id":"deepcogito/cogito-v2.1-671b","n":"Cogito v2.1 671B","f":"cogito","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1.25,"co":1.25,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"openrouter","id":"cohere/command-a","n":"Command A","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8192,"mi":"t","mo":"t","ow":1,"kn":"2024-08-31","rd":"2025-03-13","lu":"2025-03-13"},{"p":"openrouter","id":"cohere/command-r-08-2024","n":"Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.15,"co":0.6,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"openrouter","id":"cohere/command-r7b-12-2024","n":"Command R7B","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.0375,"co":0.15,"ctx":128000,"out":4000,"mi":"t","mo":"t","ow":1,"kn":"2024-06-01","rd":"2024-12-02","lu":"2024-12-02"},{"p":"openrouter","id":"cohere/command-r-plus-08-2024","n":"Command R+","f":"command-r","d":"Cohere's RAG workhorse for long-context enterprise search and tool use","ci":2.5,"co":10,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"openrouter","id":"cohere/north-mini-code:free","n":"North Mini Code (free)","f":"north","d":"Cohere coding model for practical software engineering and agentic edits","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-17","lu":"2026-06-17"},{"p":"openrouter","id":"gryphe/mythomax-l2-13b","n":"MythoMax 13B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.06,"co":0.06,"ctx":4096,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-06-30","rd":"2023-07-02","lu":"2023-07-02"},{"p":"openrouter","id":"stepfun/step-3.7-flash","n":"Step 3.7 Flash","d":"Newer StepFun flash model for faster agents, coding, and multimodal prompts","ci":0.2,"co":1.15,"cr":0.04,"ctx":256000,"out":256000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01-01","rd":"2026-05-29","lu":"2026-05-29"},{"p":"openrouter","id":"stepfun/step-3.5-flash","n":"Step 3.5 Flash","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.1,"co":0.3,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-02-13"},{"p":"openrouter","id":"nex-agi/nex-n2-mini","n":"Nex-N2-Mini","f":"agi","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.025,"co":0.1,"cr":0.0025,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-24","lu":"2026-06-24"},{"p":"openrouter","id":"nex-agi/nex-n2-pro","n":"Nex-N2-Pro","f":"agi","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.25,"co":1,"cr":0.025,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"ow":1,"rd":"2026-06-08","lu":"2026-06-08"},{"p":"openrouter","id":"undi95/remm-slerp-l2-13b","n":"ReMM SLERP 13B","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.45,"co":0.65,"ctx":6144,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-06-30","rd":"2023-07-22","lu":"2023-07-22"},{"p":"openrouter","id":"~openai/gpt-mini-latest","n":"OpenAI GPT Mini Latest","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"~openai/gpt-latest","n":"OpenAI GPT Latest","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"pit","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"~moonshotai/kimi-latest","n":"MoonshotAI Kimi Latest","f":"kimi","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.66,"co":3.41,"cr":0.14,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"relace/relace-search","n":"Relace Search","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":1,"co":3,"ctx":256000,"out":128000,"mi":"t","mo":"t","tc":1,"rd":"2025-12-08","lu":"2025-12-08"},{"p":"openrouter","id":"relace/relace-apply-3","n":"Relace Apply 3","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.85,"co":1.25,"ctx":256000,"out":128000,"mi":"t","mo":"t","rd":"2025-09-26","lu":"2025-09-26"},{"p":"openrouter","id":"ai21/jamba-large-1.7","n":"Jamba Large 1.7","f":"jamba","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":2,"co":8,"ctx":256000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-08-31","rd":"2025-08-08","lu":"2025-08-08"},{"p":"openrouter","id":"arcee-ai/coder-large","n":"Coder Large","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.5,"co":0.8,"ctx":32768,"out":32768,"mi":"t","mo":"t","kn":"2025-03-31","rd":"2025-05-05","lu":"2025-05-05"},{"p":"openrouter","id":"arcee-ai/virtuoso-large","n":"Virtuoso Large","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.75,"co":1.2,"ctx":131072,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2025-03-31","rd":"2025-05-05","lu":"2025-05-05"},{"p":"openrouter","id":"arcee-ai/trinity-large-thinking","n":"Trinity Large Thinking","f":"trinity","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":0.8,"cr":0.06,"ctx":262144,"out":80000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"openrouter","id":"arcee-ai/trinity-mini","n":"Trinity Mini","f":"trinity-mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.045,"co":0.15,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"openrouter","id":"mancer/weaver","n":"Weaver (alpha)","f":"alpha","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.75,"co":1,"ctx":8000,"out":2000,"mi":"t","mo":"t","kn":"2023-06-30","rd":"2023-08-02","lu":"2023-08-02"},{"p":"openrouter","id":"perplexity/sonar-reasoning-pro","n":"Sonar Reasoning Pro","f":"sonar-reasoning","d":"Web-grounded reasoning model for multi-step research and cited answers","ci":2,"co":8,"ctx":128000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2025-03-07","lu":"2025-03-07"},{"p":"openrouter","id":"perplexity/sonar","n":"Sonar","f":"sonar","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":1,"co":1,"ctx":127072,"out":127072,"mi":"ti","mo":"t","rd":"2025-01-27","lu":"2025-01-27"},{"p":"openrouter","id":"perplexity/sonar-pro","n":"Sonar Pro","f":"sonar-pro","d":"Advanced Sonar search model for deeper research and cited synthesis","ci":3,"co":15,"ctx":200000,"out":8000,"mi":"ti","mo":"t","rd":"2025-03-07","lu":"2025-03-07"},{"p":"openrouter","id":"perplexity/sonar-pro-search","n":"Sonar Pro Search","f":"sonar-pro","d":"Advanced Sonar search model for deeper research and cited synthesis","ci":3,"co":15,"ctx":200000,"out":8000,"mi":"ti","mo":"t","rs":1,"rd":"2025-10-30","lu":"2025-10-30"},{"p":"openrouter","id":"perplexity/sonar-deep-research","n":"Sonar Deep Research","f":"sonar-deep-research","d":"Sonar search model for current answers, retrieval, and citation-backed chat","ci":2,"co":8,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"rd":"2025-03-07","lu":"2025-03-07"},{"p":"openrouter","id":"switchpoint/router","n":"Switchpoint Router","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.85,"co":3.4,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"rd":"2025-07-11","lu":"2025-07-11"},{"p":"openrouter","id":"openrouter/bodybuilder","n":"Body Builder (beta)","d":"Preview model for early access evaluation, prototyping, and compatibility testing","ctx":128000,"out":128000,"mi":"t","mo":"t","rd":"2025-12-05","lu":"2025-12-05"},{"p":"openrouter","id":"openrouter/free","n":"Free Models Router","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0,"co":0,"ctx":200000,"out":8000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-01","lu":"2026-02-01"},{"p":"openrouter","id":"openrouter/fusion","n":"Fusion","d":"General-purpose chat model for instruction following, writing, and analysis","ctx":1000000,"out":128000,"mi":"t","mo":"t","rd":"2026-06-13","lu":"2026-06-13"},{"p":"openrouter","id":"openrouter/pareto-code","n":"Pareto Code Router","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":2000000,"out":200000,"mi":"t","mo":"t","rd":"2026-04-21","lu":"2026-04-21"},{"p":"openrouter","id":"openrouter/auto","n":"Auto Router","f":"auto","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":2000000,"out":2000000,"mi":"tiapv","mo":"ti","rs":1,"tc":1,"rd":"2023-11-08","lu":"2023-11-08"},{"p":"openrouter","id":"qwen/qwen3.5-plus-20260420","n":"Qwen3.5 Plus 2026-04-20","f":"qwen3.5","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.8,"cw":0.375,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"qwen/qwen3-next-80b-a3b-instruct:free","n":"Qwen3 Next 80B A3B Instruct (free)","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"openrouter","id":"qwen/qwen3-vl-235b-a22b-thinking","n":"Qwen3 VL 235B A22B Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.6,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-09-23","lu":"2025-09-23"},{"p":"openrouter","id":"qwen/qwen3-vl-30b-a3b-thinking","n":"Qwen3 VL 30B A3B Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.13,"co":1.56,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-10-06","lu":"2025-10-06"},{"p":"openrouter","id":"qwen/qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Hosted Qwen coder for software agents, repo edits, and long-context code","ci":0.65,"co":3.25,"cr":0.13,"cw":0.8125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"openrouter","id":"qwen/qwen-plus","n":"Qwen Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.26,"co":0.78,"cr":0.052,"cw":0.325,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-09-11"},{"p":"openrouter","id":"qwen/qwen3-coder-30b-a3b-instruct","n":"Qwen3-Coder 30B-A3B Instruct","f":"qwen","d":"Smaller Qwen coder for efficient local agents and repo-level fixes","ci":0.07,"co":0.27,"ctx":160000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"openrouter","id":"qwen/qwen3-32b","n":"Qwen3 32B","f":"qwen","d":"Dense open Qwen model for self-hosted chat, reasoning, and coding","ci":0.08,"co":0.28,"ctx":40960,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"openrouter","id":"qwen/qwen3-next-80b-a3b-instruct","n":"Qwen3-Next 80B-A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":1.1,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"openrouter","id":"qwen/qwen3-vl-8b-instruct","n":"Qwen3 VL 8B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.117,"co":0.455,"ctx":131072,"out":32768,"mi":"it","mo":"t","tc":1,"ow":1,"rd":"2025-10-14","lu":"2025-10-14"},{"p":"openrouter","id":"qwen/qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.32,"co":1.28,"cr":0.064,"cw":0.4,"ctx":1000000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"openrouter","id":"qwen/qwen3.6-35b-a3b","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.14,"co":1,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"openrouter","id":"qwen/qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":1.25,"co":3.75,"cr":0.25,"cw":1.5625,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"openrouter","id":"qwen/qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","ci":0.78,"co":3.9,"cr":0.156,"cw":0.975,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"openrouter","id":"qwen/qwen3-8b","n":"Qwen3 8B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.117,"co":0.455,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-04-28","lu":"2025-04-28"},{"p":"openrouter","id":"qwen/qwen-plus-2025-07-28","n":"Qwen Plus 0728","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.26,"co":0.78,"ctx":1000000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2025-03-31","rd":"2025-09-08","lu":"2025-09-08"},{"p":"openrouter","id":"qwen/qwen3.5-flash-02-23","n":"Qwen3.5-Flash","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.065,"co":0.26,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-25","lu":"2026-02-25"},{"p":"openrouter","id":"qwen/qwen3-coder:free","n":"Qwen3 Coder 480B A35B (free)","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":262000,"out":262000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-06-30","rd":"2025-07-23","lu":"2025-07-23"},{"p":"openrouter","id":"qwen/qwen3-30b-a3b-instruct-2507","n":"Qwen3 30B A3B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.04815,"co":0.19305,"ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-06-30","rd":"2025-07-29","lu":"2025-07-29"},{"p":"openrouter","id":"qwen/qwen-2.5-coder-32b-instruct","n":"Qwen2.5 Coder 32B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.66,"co":1,"ctx":32768,"out":32768,"mi":"t","mo":"t","ow":1,"kn":"2024-06-30","rd":"2024-11-11","lu":"2024-11-11"},{"p":"openrouter","id":"qwen/qwen3-next-80b-a3b-thinking","n":"Qwen3-Next 80B-A3B (Thinking)","f":"qwen","d":"Efficient Qwen thinking model for local reasoning, math, and coding agents","ci":0.0975,"co":0.78,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"openrouter","id":"qwen/qwen3-235b-a22b-thinking-2507","n":"Qwen3 235B A22B Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.1495,"co":1.495,"ctx":131072,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06-30","rd":"2025-07-25","lu":"2025-07-25"},{"p":"openrouter","id":"qwen/qwen3-vl-32b-instruct","n":"Qwen3 VL 32B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.104,"co":0.416,"ctx":131072,"out":32768,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2025-10-23","lu":"2025-10-23"},{"p":"openrouter","id":"qwen/qwen3-coder","n":"Qwen3 Coder 480B A35B","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.22,"co":1.8,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-06-30","rd":"2025-07-23","lu":"2025-07-23"},{"p":"openrouter","id":"qwen/qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1875,"co":1.125,"cw":0.234375,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"openrouter","id":"qwen/qwen3.5-plus-02-15","n":"Qwen3.5 Plus 2026-02-15","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":1.56,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"openrouter","id":"qwen/qwen-2.5-7b-instruct","n":"Qwen2.5 7B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.04,"co":0.1,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-30","rd":"2024-10-16","lu":"2024-10-16"},{"p":"openrouter","id":"qwen/qwen3-vl-8b-thinking","n":"Qwen3 VL 8B Thinking","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.117,"co":1.365,"ctx":131072,"out":32768,"mi":"it","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-14","lu":"2025-10-14"},{"p":"openrouter","id":"qwen/qwen3-max-thinking","n":"Qwen3 Max Thinking","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.78,"co":3.9,"ctx":262144,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-09","lu":"2026-02-09"},{"p":"openrouter","id":"qwen/qwen3-30b-a3b-thinking-2507","n":"Qwen3 30B A3B Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.13,"co":1.56,"ctx":81920,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-06-30","rd":"2025-08-28","lu":"2025-08-28"},{"p":"openrouter","id":"qwen/qwen2.5-vl-72b-instruct","n":"Qwen2.5 VL 72B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.8,"co":1,"cr":0.4,"ctx":128000,"out":128000,"mi":"ti","mo":"t","ow":1,"kn":"2024-06-30","rd":"2025-02-01","lu":"2025-02-01"},{"p":"openrouter","id":"qwen/qwen3.5-27b","n":"Qwen3.5 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.195,"co":1.56,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"openrouter","id":"qwen/qwen3-235b-a22b","n":"Qwen3 235B-A22B","f":"qwen","d":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","ci":0.455,"co":1.82,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"openrouter","id":"qwen/qwen-2.5-72b-instruct","n":"Qwen2.5 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.36,"co":0.4,"ctx":32768,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-30","rd":"2024-09-19","lu":"2024-09-19"},{"p":"openrouter","id":"qwen/qwen-plus-2025-07-28:thinking","n":"Qwen Plus 0728 (thinking)","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.26,"co":0.78,"cw":0.325,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-09-08","lu":"2025-09-08"},{"p":"openrouter","id":"qwen/qwen3-coder-next","n":"Qwen3 Coder Next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.11,"co":0.8,"cr":0.07,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-04","lu":"2026-02-04"},{"p":"openrouter","id":"qwen/qwen3.6-27b","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.285,"co":2.4,"cr":0.15,"ctx":262140,"out":262140,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"openrouter","id":"qwen/qwen3.5-35b-a3b","n":"Qwen3.5 35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.14,"co":1,"cr":0.05,"ctx":262144,"out":81920,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"openrouter","id":"qwen/qwen3.5-9b","n":"Qwen3.5 9B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.15,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"openrouter","id":"qwen/qwen3.5-397b-a17b","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.385,"co":2.45,"cr":0.111,"ctx":131072,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"openrouter","id":"qwen/qwen3-vl-30b-a3b-instruct","n":"Qwen3 VL 30B A3B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.13,"co":0.52,"ctx":131072,"out":32768,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-10-06","lu":"2025-10-06"},{"p":"openrouter","id":"qwen/qwen3-235b-a22b-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.1,"ctx":262144,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-06-30","rd":"2025-07-21","lu":"2025-07-21"},{"p":"openrouter","id":"qwen/qwen3-coder-flash","n":"Qwen3 Coder Flash","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.195,"co":0.975,"cr":0.039,"cw":0.24375,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"openrouter","id":"qwen/qwen3-14b","n":"Qwen3 14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.24,"ctx":40960,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-04-28","lu":"2025-04-28"},{"p":"openrouter","id":"qwen/qwen3-vl-235b-a22b-instruct","n":"Qwen3 VL 235B A22B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.88,"cr":0.11,"ctx":262144,"out":16384,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-09-23","lu":"2025-09-23"},{"p":"openrouter","id":"qwen/qwen3-30b-a3b","n":"Qwen3 30B A3B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.12,"co":0.5,"ctx":40960,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-04-28","lu":"2025-04-28"},{"p":"openrouter","id":"qwen/qwen3.6-max-preview","n":"Qwen3.6 Max Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.04,"co":6.24,"cw":1.3,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-20","lu":"2026-04-20"},{"p":"openrouter","id":"qwen/qwen3.5-122b-a10b","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.26,"co":2.08,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"openrouter","id":"qwen/qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.325,"co":1.95,"cw":0.40625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"openrouter","id":"amazon/nova-lite-v1","n":"Nova Lite 1.0","f":"nova-lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.06,"co":0.24,"ctx":300000,"out":5120,"mi":"ti","mo":"t","tc":1,"kn":"2024-10-31","rd":"2024-12-05","lu":"2024-12-05"},{"p":"openrouter","id":"amazon/nova-premier-v1","n":"Nova Premier 1.0","f":"nova","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":2.5,"co":12.5,"cr":0.625,"ctx":1000000,"out":32000,"mi":"ti","mo":"t","tc":1,"rd":"2025-10-31","lu":"2025-10-31"},{"p":"openrouter","id":"amazon/nova-pro-v1","n":"Nova Pro 1.0","f":"nova-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.8,"co":3.2,"ctx":300000,"out":5120,"mi":"ti","mo":"t","tc":1,"kn":"2024-10-31","rd":"2024-12-05","lu":"2024-12-05"},{"p":"openrouter","id":"amazon/nova-micro-v1","n":"Nova Micro 1.0","f":"nova-micro","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.035,"co":0.14,"ctx":128000,"out":5120,"mi":"t","mo":"t","tc":1,"kn":"2024-10-31","rd":"2024-12-05","lu":"2024-12-05"},{"p":"openrouter","id":"amazon/nova-2-lite-v1","n":"Nova 2 Lite","f":"nova","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.3,"co":2.5,"ctx":1000000,"out":65535,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"openrouter","id":"aion-labs/aion-rp-llama-3.1-8b","n":"Aion-RP 1.0 (8B)","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.8,"co":1.6,"ctx":32768,"out":32768,"mi":"t","mo":"t","kn":"2023-12-31","rd":"2025-02-04","lu":"2025-02-04"},{"p":"openrouter","id":"aion-labs/aion-1.0-mini","n":"Aion-1.0-Mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.7,"co":1.4,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-02-04","lu":"2025-02-04"},{"p":"openrouter","id":"aion-labs/aion-2.0","n":"Aion-2.0","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.8,"co":1.6,"cr":0.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"openrouter","id":"aion-labs/aion-1.0","n":"Aion-1.0","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":4,"co":8,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-02-04","lu":"2025-02-04"},{"p":"openrouter","id":"inflection/inflection-3-pi","n":"Inflection 3 Pi","d":"General-purpose chat model for instruction following, writing, and analysis","ci":2.5,"co":10,"ctx":8000,"out":1024,"mi":"t","mo":"t","kn":"2024-10-31","rd":"2024-10-11","lu":"2024-10-11"},{"p":"openrouter","id":"inflection/inflection-3-productivity","n":"Inflection 3 Productivity","d":"General-purpose chat model for instruction following, writing, and analysis","ci":2.5,"co":10,"ctx":8000,"out":1024,"mi":"t","mo":"t","kn":"2024-10-31","rd":"2024-10-11","lu":"2024-10-11"},{"p":"openrouter","id":"sao10k/l3.1-euryale-70b","n":"Llama 3.1 Euryale 70B v2.2","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.85,"co":0.85,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12-31","rd":"2024-08-28","lu":"2024-08-28"},{"p":"openrouter","id":"sao10k/l3.3-euryale-70b","n":"Llama 3.3 Euryale 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.65,"co":0.75,"ctx":131072,"out":16384,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-12-18","lu":"2024-12-18"},{"p":"openrouter","id":"sao10k/l3-lunaris-8b","n":"Llama 3 8B Lunaris","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.04,"co":0.05,"ctx":8192,"out":16384,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-08-13","lu":"2024-08-13"},{"p":"openrouter","id":"sao10k/l3.1-70b-hanami-x1","n":"Llama 3.1 70B Hanami x1","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":3,"co":3,"ctx":16000,"out":16000,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2025-01-08","lu":"2025-01-08"},{"p":"openrouter","id":"upstage/solar-pro-3","n":"Solar Pro 3","f":"solar-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.15,"co":0.6,"cr":0.015,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-01-27","lu":"2026-01-27"},{"p":"openrouter","id":"allenai/olmo-3-32b-think","n":"Olmo 3 32B Think","f":"allenai","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.15,"co":0.5,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2025-11-21","lu":"2025-11-21"},{"p":"openrouter","id":"deepseek/deepseek-r1-0528","n":"R1 0528","f":"deepseek","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.5,"co":2.15,"cr":0.35,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-05-28","lu":"2025-05-28"},{"p":"openrouter","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.09,"co":0.18,"cr":0.018,"ctx":1048576,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"openrouter","id":"deepseek/deepseek-v3.1-terminus","n":"DeepSeek V3.1 Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.95,"cr":0.13,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-09-22","lu":"2025-09-22"},{"p":"openrouter","id":"deepseek/deepseek-r1-distill-llama-70b","n":"R1 Distill Llama 70B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.8,"co":0.8,"ctx":8192,"out":8192,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07-31","rd":"2025-01-23","lu":"2025-01-23"},{"p":"openrouter","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"openrouter","id":"deepseek/deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"Classic open reasoning model for transparent math, coding, and deliberate problem solving","ci":0.7,"co":2.5,"ctx":64000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-05-29"},{"p":"openrouter","id":"deepseek/deepseek-v3.2-exp","n":"DeepSeek V3.2 Exp","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":0.41,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"openrouter","id":"deepseek/deepseek-chat-v3-0324","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.24,"co":0.9,"cr":0.135,"ctx":163840,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07-31","rd":"2025-03-24","lu":"2025-03-24"},{"p":"openrouter","id":"deepseek/deepseek-chat","n":"DeepSeek Chat","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2002,"co":0.8001,"ctx":128000,"out":16000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"openrouter","id":"deepseek/deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2288,"co":0.3432,"cr":0.02288,"ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"openrouter","id":"deepseek/deepseek-chat-v3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.21,"co":0.79,"cr":0.13,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-03-31","rd":"2025-08-21","lu":"2025-08-21"},{"p":"openrouter","id":"minimax/minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.12,"co":0.48,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"openrouter","id":"minimax/minimax-m2.1","n":"MiniMax-M2.1","f":"minimax","d":"Earlier MiniMax agent model for practical coding and productivity tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"openrouter","id":"minimax/minimax-01","n":"MiniMax-01","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.2,"co":1.1,"ctx":1000192,"out":1000192,"mi":"ti","mo":"t","ow":1,"kn":"2024-03-31","rd":"2025-01-15","lu":"2025-01-15"},{"p":"openrouter","id":"minimax/minimax-m3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.3,"co":1.2,"cr":0.06,"ctx":524288,"out":512000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"openrouter","id":"minimax/minimax-m2","n":"MiniMax-M2","f":"minimax","d":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","ci":0.255,"co":1.02,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"openrouter","id":"minimax/minimax-m2-her","n":"MiniMax M2-her","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.03,"ctx":65536,"out":2048,"mi":"t","mo":"t","rd":"2026-01-23","lu":"2026-01-23"},{"p":"openrouter","id":"minimax/minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.18,"co":0.72,"ctx":196608,"out":196608,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"openrouter","id":"minimax/minimax-m1","n":"MiniMax M1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.4,"co":2.2,"ctx":1000000,"out":40000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-06-30","rd":"2025-06-17","lu":"2025-06-17"},{"p":"openrouter","id":"kwaipilot/kat-coder-pro-v2","n":"KAT-Coder-Pro V2","f":"kat-coder","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":256000,"out":80000,"mi":"t","mo":"t","tc":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"openrouter","id":"nousresearch/hermes-3-llama-3.1-405b:free","n":"Hermes 3 405B Instruct (free)","f":"hermes","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-08-16","lu":"2024-08-16"},{"p":"openrouter","id":"nousresearch/hermes-4-405b","n":"Hermes 4 405B","f":"hermes","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1,"co":3,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-08-31","rd":"2025-08-26","lu":"2025-08-26"},{"p":"openrouter","id":"nousresearch/hermes-3-llama-3.1-70b","n":"Hermes 3 70B Instruct","f":"nousresearch","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.7,"co":0.7,"ctx":131072,"out":16384,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-08-18","lu":"2024-08-18"},{"p":"openrouter","id":"nousresearch/hermes-3-llama-3.1-405b","n":"Hermes 3 405B Instruct","f":"nousresearch","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":1,"co":1,"ctx":131072,"out":16384,"mi":"t","mo":"t","ow":1,"kn":"2023-12-31","rd":"2024-08-16","lu":"2024-08-16"},{"p":"openrouter","id":"nousresearch/hermes-4-70b","n":"Hermes 4 70B","f":"hermes","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.13,"co":0.4,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-08-31","rd":"2025-08-26","lu":"2025-08-26"},{"p":"jiekou","id":"o3","n":"o3","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":10,"co":40,"ctx":131072,"out":131072,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"grok-code-fast-1","n":"grok-code-fast-1","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.18,"co":1.35,"ctx":256000,"out":256000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.2-pro","n":"gpt-5.2-pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":18.9,"co":151.2,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-pro","n":"gemini-2.5-pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.125,"co":9,"ctx":1048576,"out":65535,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"claude-haiku-4-5-20251001","n":"claude-haiku-4-5-20251001","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.9,"co":4.5,"ctx":20000,"out":64000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5-pro","n":"gpt-5-pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":13.5,"co":108,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-flash","n":"gemini-2.5-flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.27,"co":2.25,"ctx":1048576,"out":65535,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"o4-mini","n":"o4-mini","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-flash-lite-preview-09-2025","n":"gemini-2.5-flash-lite-preview-09-2025","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.09,"co":0.36,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"claude-opus-4-1-20250805","n":"claude-opus-4-1-20250805","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":13.5,"co":67.5,"ctx":200000,"out":32000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"grok-4-1-fast-non-reasoning","n":"grok-4-1-fast-non-reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.18,"co":0.45,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5-chat-latest","n":"gpt-5-chat-latest","f":"gpt","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.125,"co":9,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"claude-opus-4-5-20251101","n":"claude-opus-4-5-20251101","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.5,"co":22.5,"ctx":200000,"out":65536,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.1-codex","n":"gpt-5.1-codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.125,"co":9,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.1-codex-max","n":"gpt-5.1-codex-max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.125,"co":9,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"claude-opus-4-20250514","n":"claude-opus-4-20250514","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":13.5,"co":67.5,"ctx":200000,"out":32000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"o3-mini","n":"o3-mini","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":131072,"out":131072,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.2","n":"gpt-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.575,"co":12.6,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"grok-4-0709","n":"grok-4-0709","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2.7,"co":13.5,"ctx":256000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-flash-lite","n":"gemini-2.5-flash-lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.09,"co":0.36,"ctx":1048576,"out":65535,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"claude-sonnet-4-20250514","n":"claude-sonnet-4-20250514","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.7,"co":13.5,"ctx":200000,"out":64000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.1-codex-mini","n":"gpt-5.1-codex-mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.225,"co":1.8,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"grok-4-fast-reasoning","n":"grok-4-fast-reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.18,"co":0.45,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"claude-opus-4-6","n":"claude-opus-4-6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02","lu":"2026-02"},{"p":"jiekou","id":"claude-sonnet-4-5-20250929","n":"claude-sonnet-4-5-20250929","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.7,"co":13.5,"ctx":200000,"out":64000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"grok-4-1-fast-reasoning","n":"grok-4-1-fast-reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.18,"co":0.45,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-3-pro-preview","n":"gemini-3-pro-preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.8,"co":10.8,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5-mini","n":"gpt-5-mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.225,"co":1.8,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5-nano","n":"gpt-5-nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.045,"co":0.36,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"grok-4-fast-non-reasoning","n":"grok-4-fast-non-reasoning","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.18,"co":0.45,"ctx":2000000,"out":2000000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-3-flash-preview","n":"gemini-3-flash-preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"ctx":1048576,"out":65536,"mi":"tiva","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-flash-preview-05-20","n":"gemini-2.5-flash-preview-05-20","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.135,"co":3.15,"ctx":1048576,"out":200000,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-flash-lite-preview-06-17","n":"gemini-2.5-flash-lite-preview-06-17","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.09,"co":0.36,"ctx":1048576,"out":65535,"mi":"tvia","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gemini-2.5-pro-preview-06-05","n":"gemini-2.5-pro-preview-06-05","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.125,"co":9,"ctx":1048576,"out":200000,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5-codex","n":"gpt-5-codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.125,"co":9,"ctx":400000,"out":128000,"mi":"ti","mo":"t","tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.2-codex","n":"gpt-5.2-codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"gpt-5.1","n":"gpt-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.125,"co":9,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02","lu":"2026-02"},{"p":"jiekou","id":"moonshotai/kimi-k2-instruct","n":"Kimi K2 Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.57,"co":2.3,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"moonshotai/kimi-k2-0905","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.6,"co":2.5,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"minimaxai/minimax-m1-80k","n":"MiniMax M1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.55,"co":2.2,"ctx":1000000,"out":40000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"baidu/ernie-4.5-vl-424b-a47b","n":"ERNIE 4.5 VL 424B A47B","f":"ernie","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.42,"co":1.25,"ctx":123000,"out":16000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"baidu/ernie-4.5-300b-a47b-paddle","n":"ERNIE 4.5 300B A47B","f":"ernie","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.28,"co":1.1,"ctx":123000,"out":12000,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"xiaomimimo/mimo-v2-flash","n":"XiaomiMiMo/MiMo-V2-Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0,"co":0,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"zai-org/glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"zai-org/glm-4.5v","n":"GLM 4.5V","f":"glmv","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.8,"ctx":65536,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"zai-org/glm-4.5","n":"GLM-4.5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"zai-org/glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-235b-a22b-instruct-2507","n":"Qwen3 235B A22B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":0.8,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-next-80b-a3b-instruct","n":"Qwen3 Next 80B A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":1.5,"ctx":65536,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-next-80b-a3b-thinking","n":"Qwen3 Next 80B A3B Thinking","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.15,"co":1.5,"ctx":65536,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-235b-a22b-thinking-2507","n":"Qwen3 235B A22b Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.3,"co":3,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-32b-fp8","n":"Qwen3 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.45,"ctx":40960,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-30b-a3b-fp8","n":"Qwen3 30B A3B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.09,"co":0.45,"ctx":40960,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-coder-next","n":"qwen/qwen3-coder-next","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.2,"co":1.5,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02","lu":"2026-02"},{"p":"jiekou","id":"qwen/qwen3-coder-480b-a35b-instruct","n":"Qwen3 Coder 480B A35B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.29,"co":1.2,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"qwen/qwen3-235b-a22b-fp8","n":"Qwen3 235B A22B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.2,"co":0.8,"ctx":40960,"out":20000,"mi":"t","mo":"t","rs":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"deepseek/deepseek-r1-0528","n":"DeepSeek R1 0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.7,"co":2.5,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"deepseek/deepseek-v3-0324","n":"DeepSeek V3 0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.28,"co":1.14,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"deepseek/deepseek-v3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27,"co":1,"ctx":163840,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"jiekou","id":"minimax/minimax-m2.1","n":"Minimax M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01","lu":"2026-01"},{"p":"nova","id":"nova-2-pro-v1","n":"Nova 2 Pro","f":"nova-pro","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0,"co":0,"ctx":1000000,"out":64000,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2025-12-03","lu":"2026-01-03"},{"p":"nova","id":"nova-2-lite-v1","n":"Nova 2 Lite","f":"nova-lite","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0,"co":0,"ctx":1000000,"out":64000,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"alibaba-token-plan","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0,"co":0,"cr":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"alibaba-token-plan","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"alibaba-token-plan","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"alibaba-token-plan","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"alibaba-token-plan","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"alibaba-token-plan","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0,"co":0,"cr":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"alibaba-token-plan","id":"wan2.7-image-pro","n":"Wan2.7 Image Pro","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-05-29","lu":"2026-05-29"},{"p":"alibaba-token-plan","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"alibaba-token-plan","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"alibaba-token-plan","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"alibaba-token-plan","id":"qwen-image-2.0","n":"Qwen Image 2.0","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-03-03","lu":"2026-03-03"},{"p":"alibaba-token-plan","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0,"co":0,"cr":0,"cw":0,"ctx":196608,"out":24576,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-token-plan","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0,"co":0,"cr":0,"cw":0,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"alibaba-token-plan","id":"qwen-image-2.0-pro","n":"Qwen Image 2.0 Pro","f":"qwen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-03-03","lu":"2026-03-03"},{"p":"alibaba-token-plan","id":"wan2.7-image","n":"Wan2.7 Image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0,"co":0,"ctx":8192,"out":0,"mi":"t","mo":"i","rd":"2026-05-29","lu":"2026-05-29"},{"p":"alibaba-token-plan","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":202752,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"alibaba-token-plan","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0,"co":0,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2025-12-03","lu":"2025-12-05"},{"p":"alibaba-token-plan","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"alibaba","id":"qwen3-omni-flash","n":"Qwen3-Omni Flash","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.43,"co":1.66,"ctx":65536,"out":16384,"mi":"tiav","mo":"ta","rs":1,"tc":1,"kn":"2024-04","rd":"2025-09-15","lu":"2025-09-15"},{"p":"alibaba","id":"qwen3-coder-plus","n":"Qwen3 Coder Plus","f":"qwen","d":"Hosted Qwen coder for software agents, repo edits, and long-context code","ci":1,"co":5,"ctx":1048576,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"alibaba","id":"qwen-plus","n":"Qwen Plus","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.4,"co":1.2,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-09-11"},{"p":"alibaba","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3-Coder 30B-A3B Instruct","f":"qwen","d":"Smaller Qwen coder for efficient local agents and repo-level fixes","ci":0.45,"co":2.25,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen3-omni-flash-realtime","n":"Qwen3-Omni Flash Realtime","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.52,"co":1.99,"ctx":65536,"out":16384,"mi":"tiav","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-09-15","lu":"2025-09-15"},{"p":"alibaba","id":"qwen3-32b","n":"Qwen3 32B","f":"qwen","d":"Dense open Qwen model for self-hosted chat, reasoning, and coding","ci":0.7,"co":2.8,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen-omni-turbo-realtime","n":"Qwen-Omni Turbo Realtime","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.27,"co":1.07,"ctx":32768,"out":2048,"mi":"tia","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-05-08","lu":"2025-05-08"},{"p":"alibaba","id":"qwen-plus-character-ja","n":"Qwen Plus Character (Japanese)","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.5,"co":1.4,"ctx":8192,"out":512,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01","lu":"2024-01"},{"p":"alibaba","id":"qwen3-next-80b-a3b-instruct","n":"Qwen3-Next 80B-A3B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.5,"co":2,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"alibaba","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-04"},{"p":"alibaba","id":"qwen3.6-35b-a3b","n":"Qwen3.6 35B-A3B","f":"qwen","d":"Open multimodal Qwen MoE for local agents that need vision, audio, and code","ci":0.248,"co":1.485,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"alibaba","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":2.5,"co":7.5,"cr":0.5,"cw":3.125,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"alibaba","id":"qwen3-max","n":"Qwen3 Max","f":"qwen","d":"Flagship Qwen3 model for coding agents, complex reasoning, and tool use","ci":1.2,"co":6,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"alibaba","id":"qwen2-5-omni-7b","n":"Qwen2.5-Omni 7B","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.1,"co":0.4,"ctx":32768,"out":2048,"mi":"tiav","mo":"ta","tc":1,"ow":1,"kn":"2024-04","rd":"2024-12","lu":"2024-12"},{"p":"alibaba","id":"qwen3-8b","n":"Qwen3 8B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.18,"co":0.7,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen2-5-14b-instruct","n":"Qwen2.5 14B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.35,"co":1.4,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba","id":"qwen3-next-80b-a3b-thinking","n":"Qwen3-Next 80B-A3B (Thinking)","f":"qwen","d":"Efficient Qwen thinking model for local reasoning, math, and coding agents","ci":0.5,"co":6,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-09","lu":"2025-09"},{"p":"alibaba","id":"qvq-max","n":"QVQ Max","f":"qvq","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.2,"co":4.8,"ctx":131072,"out":8192,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-03-25","lu":"2025-03-25"},{"p":"alibaba","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1875,"co":1.125,"cw":0.234375,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-27","lu":"2026-04-27"},{"p":"alibaba","id":"qwen2-5-vl-72b-instruct","n":"Qwen2.5-VL 72B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":2.8,"co":8.4,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba","id":"qwen3-vl-plus","n":"Qwen3-VL Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1.6,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2025-09-23","lu":"2025-09-23"},{"p":"alibaba","id":"qwen-vl-ocr","n":"Qwen-VL OCR","f":"qwen","d":"OCR model for extracting structured text from documents and screenshots","ci":0.72,"co":0.72,"ctx":34096,"out":4096,"mi":"ti","mo":"t","kn":"2024-04","rd":"2024-10-28","lu":"2025-04-13"},{"p":"alibaba","id":"qwen-mt-turbo","n":"Qwen-MT Turbo","f":"qwen","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":0.16,"co":0.49,"ctx":16384,"out":8192,"mi":"t","mo":"t","kn":"2024-04","rd":"2025-01","lu":"2025-01"},{"p":"alibaba","id":"qwen-mt-plus","n":"Qwen-MT Plus","f":"qwen","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":2.46,"co":7.37,"ctx":16384,"out":8192,"mi":"t","mo":"t","kn":"2024-04","rd":"2025-01","lu":"2025-01"},{"p":"alibaba","id":"qwen3.5-plus","n":"Qwen3.5 Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":2.4,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-02-16"},{"p":"alibaba","id":"qwen-omni-turbo","n":"Qwen-Omni Turbo","f":"qwen","d":"Qwen omni model for text, vision, audio, and multimodal agent tasks","ci":0.07,"co":0.27,"ctx":32768,"out":2048,"mi":"tiav","mo":"ta","tc":1,"kn":"2024-04","rd":"2025-01-19","lu":"2025-03-26"},{"p":"alibaba","id":"qwen2-5-72b-instruct","n":"Qwen2.5 72B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":1.4,"co":5.6,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba","id":"qwen-flash","n":"Qwen Flash","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.05,"co":0.4,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"alibaba","id":"qwen3-vl-235b-a22b","n":"Qwen3-VL 235B-A22B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.7,"co":2.8,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen3-vl-30b-a3b","n":"Qwen3-VL 30B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":0.8,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen-vl-max","n":"Qwen-VL Max","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.8,"co":3.2,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-04-08","lu":"2025-08-13"},{"p":"alibaba","id":"qwen3.5-27b","n":"Qwen3.5 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":2.4,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"alibaba","id":"qwen-max","n":"Qwen Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.6,"co":6.4,"ctx":32768,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2024-04","rd":"2024-04-03","lu":"2025-01-25"},{"p":"alibaba","id":"qwen3-235b-a22b","n":"Qwen3 235B-A22B","f":"qwen","d":"Large open Qwen MoE for multilingual reasoning, coding, and tool use","ci":0.7,"co":2.8,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen3-livetranslate-flash-realtime","n":"Qwen3-LiveTranslate Flash Realtime","f":"qwen","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":10,"co":10,"ctx":53248,"out":4096,"mi":"tiav","mo":"ta","kn":"2024-04","rd":"2025-09-22","lu":"2025-09-22"},{"p":"alibaba","id":"qwen3.6-27b","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"alibaba","id":"qwen3.5-35b-a3b","n":"Qwen3.5 35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.25,"co":2,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"alibaba","id":"qwen3-coder-480b-a35b-instruct","n":"Qwen3-Coder 480B-A35B Instruct","f":"qwen","d":"Open Qwen coding heavyweight for repository reasoning and agentic engineering","ci":1.5,"co":7.5,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwq-plus","n":"QwQ Plus","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.8,"co":2.4,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-03-05","lu":"2025-03-05"},{"p":"alibaba","id":"qwen2-5-32b-instruct","n":"Qwen2.5 32B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.7,"co":2.8,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba","id":"qwen3.5-397b-a17b","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.6,"co":3.6,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"alibaba","id":"qwen3-coder-flash","n":"Qwen3 Coder Flash","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.3,"co":1.5,"ctx":1000000,"out":65536,"mi":"t","mo":"t","tc":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"alibaba","id":"qwen3-14b","n":"Qwen3 14B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.35,"co":1.4,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"alibaba","id":"qwen3-asr-flash","n":"Qwen3-ASR Flash","f":"qwen","d":"Speech transcription model for accurate audio-to-text and captioning workflows","ci":0.035,"co":0.035,"ctx":53248,"out":4096,"mi":"a","mo":"t","kn":"2024-04","rd":"2025-09-08","lu":"2025-09-08"},{"p":"alibaba","id":"qwen-turbo","n":"Qwen Turbo","f":"qwen","d":"Efficient Qwen model for fast chat, extraction, and high-volume workloads","ci":0.05,"co":0.2,"ctx":1000000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2024-11-01","lu":"2025-04-28"},{"p":"alibaba","id":"qwen2-5-7b-instruct","n":"Qwen2.5 7B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.175,"co":0.7,"ctx":131072,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba","id":"qwen2-5-vl-7b-instruct","n":"Qwen2.5-VL 7B Instruct","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.35,"co":1.05,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-04","rd":"2024-09","lu":"2024-09"},{"p":"alibaba","id":"qwen3.6-max-preview","n":"Qwen3.6 Max Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.3,"co":7.8,"cr":0.13,"cw":1.625,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-20","lu":"2026-04-20"},{"p":"alibaba","id":"qwen3.5-122b-a10b","n":"Qwen3.5 122B-A10B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":3.2,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-23","lu":"2026-02-23"},{"p":"alibaba","id":"qwen-vl-plus","n":"Qwen-VL Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.21,"co":0.63,"ctx":131072,"out":8192,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2024-01-25","lu":"2025-08-15"},{"p":"alibaba","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen","d":"Earlier Qwen multimodal workhorse for million-token agent and document tasks","ci":0.5,"co":3,"cr":0.05,"cw":0.625,"ctx":1000000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"databricks","id":"databricks-claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"databricks","id":"databricks-gpt-5-4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"databricks","id":"databricks-gemini-3-flash","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"databricks","id":"databricks-claude-opus-4-5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"databricks","id":"databricks-gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"databricks","id":"databricks-gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"databricks","id":"databricks-gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"databricks","id":"databricks-gemini-2-5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"databricks","id":"databricks-gemini-3-1-pro","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"databricks","id":"databricks-gemini-2-5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"databricks","id":"databricks-claude-sonnet-4","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"databricks","id":"databricks-claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"databricks","id":"databricks-gpt-5-4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"databricks","id":"databricks-claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"databricks","id":"databricks-gpt-5-1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"databricks","id":"databricks-claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"databricks","id":"databricks-claude-opus-4-1","n":"Claude Opus 4.1 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"databricks","id":"databricks-gpt-5-5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"databricks","id":"databricks-gemini-3-1-flash-lite","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"databricks","id":"databricks-gemini-3-pro","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"databricks","id":"databricks-claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"databricks","id":"databricks-gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"databricks","id":"databricks-gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.072,"co":0.28,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"databricks","id":"databricks-gpt-5-2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"databricks","id":"databricks-gpt-5-4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"crof","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.12,"co":0.21,"cr":0.003,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"crof","id":"minimax-m2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.11,"co":0.95,"cr":0.02,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"crof","id":"greg-2-ultra","n":"Greg 2 Ultra","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":3,"co":10,"cr":0.5,"ctx":229376,"out":229376,"mi":"t","mo":"t","rd":"2026-06-14","lu":"2026-06-14"},{"p":"crof","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Mature GLM model for dependable coding, reasoning, and structured agent tasks","ci":0.25,"co":1.1,"cr":0.05,"cw":0,"ctx":202752,"out":202752,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"crof","id":"deepseek-v4-pro-lightning","n":"DeepSeek V4 Pro Lightning","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.8,"co":1.6,"cr":0.02,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"crof","id":"greg-rp","n":"Greg (Roleplay)","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1,"co":0.3,"cr":0.02,"ctx":229376,"out":229376,"mi":"t","mo":"t","rd":"2026-01-27","lu":"2026-01-27"},{"p":"crof","id":"gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.1,"co":0.3,"cr":0.02,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"crof","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.55,"co":2.25,"cr":0.05,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"crof","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.45,"co":2.15,"cr":0.08,"cw":0,"ctx":202752,"out":202752,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"crof","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.35,"co":0.8,"cr":0.003,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"crof","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0.5,"co":2.2,"cr":0.08,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"crof","id":"greg-2-super","n":"Greg 2 Super","d":"General-purpose chat model for instruction following, writing, and analysis","ci":1.5,"co":5,"cr":0.25,"ctx":229376,"out":229376,"mi":"t","mo":"t","rd":"2026-06-14","lu":"2026-06-14"},{"p":"crof","id":"kimi-k2.5-lightning","n":"Kimi K2.5 (Lightning)","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":1,"co":3,"cr":0.2,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-06","lu":"2026-02-06"},{"p":"crof","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.35,"co":1.7,"cr":0.07,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"crof","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.5,"co":1.99,"cr":0.05,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"crof","id":"qwen3.6-27b","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.2,"co":1.5,"cr":0.04,"ctx":262144,"out":262144,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"crof","id":"qwen3.5-9b","n":"Qwen3.5 9B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.04,"co":0.15,"cr":0.008,"ctx":262144,"out":262144,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-13","lu":"2026-03-13"},{"p":"crof","id":"qwen3.5-397b-a17b","n":"Qwen3.5 397B-A17B","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.35,"co":1.75,"cr":0.07,"ctx":262144,"out":262144,"mi":"tiva","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"crof","id":"glm-4.7-flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0.04,"co":0.3,"cr":0.008,"cw":0,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"crof","id":"mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.4,"co":0.8,"cr":0.003,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"crof","id":"glm-5","n":"GLM-5","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.48,"co":1.9,"cr":0.1,"cw":0,"ctx":202752,"out":202752,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"crof","id":"deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.18,"co":0.35,"cr":0.04,"ctx":163840,"out":163840,"mi":"t","mo":"t","tc":1,"rd":"2025-07-22","lu":"2025-07-22"},{"p":"crof","id":"greg-1-mini","n":"Greg 1 Mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.07,"co":0.15,"cr":0.01,"ctx":229376,"out":229376,"mi":"t","mo":"t","rd":"2026-01-27","lu":"2026-01-27"},{"p":"fastrouter","id":"wanx/wan-v2-6","n":"Wan 2.6","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":400000,"out":0,"mi":"ti","mo":"v","ow":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"fastrouter","id":"moonshotai/kimi-k2","n":"Kimi K2","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.55,"co":2.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2025-07-11","lu":"2025-07-11"},{"p":"fastrouter","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.75,"co":3.5,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"fastrouter","id":"google/imagen-4.0-fast","n":"Imagen 4 Fast","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","rd":"2025-05-20","lu":"2025-05-20"},{"p":"fastrouter","id":"google/gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.31,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"fastrouter","id":"google/gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.0375,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"fastrouter","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"fastrouter","id":"google/veo3.1-lite","n":"Veo 3.1 Lite","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":400000,"out":0,"mi":"ti","mo":"v","rd":"2026-05-01","lu":"2026-05-01"},{"p":"fastrouter","id":"google/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.13,"co":0.38,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"fastrouter","id":"google/veo3.1","n":"Veo 3.1","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":400000,"out":0,"mi":"ti","mo":"v","rd":"2026-05-01","lu":"2026-05-01"},{"p":"fastrouter","id":"google/imagen-4.0-ultra","n":"Imagen 4 Ultra","f":"imagen","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":480,"out":0,"mi":"t","mo":"i","rd":"2025-05-20","lu":"2025-05-20"},{"p":"fastrouter","id":"google/gemini-3-pro-image-preview","n":"Nano Banana Pro","f":"gemini-pro","d":"Nano Banana Pro for higher-fidelity image generation and design-heavy edits","ci":2,"co":12,"ctx":65536,"out":32768,"mi":"ti","mo":"ti","rs":1,"kn":"2025-01","rd":"2025-11-20","lu":"2025-11-20"},{"p":"fastrouter","id":"google/gemini-3.1-flash-image-preview","n":"Nano Banana 2","f":"gemini-flash","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.5,"co":3,"ctx":65536,"out":65536,"mi":"tip","mo":"ti","rs":1,"kn":"2025-01","rd":"2026-02-26","lu":"2026-02-26"},{"p":"fastrouter","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"fastrouter","id":"google/veo3.1-fast","n":"Veo 3.1 Fast","f":"veo","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":400000,"out":0,"mi":"ti","mo":"v","rd":"2026-05-01","lu":"2026-05-01"},{"p":"fastrouter","id":"x-ai/grok-4.3","n":"Grok 4.3","f":"grok","d":"xAI's default Grok for chat, coding, agentic tools, and lower hallucination risk","ci":1.25,"co":2.5,"ctx":1000000,"out":30000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-17","lu":"2026-04-17"},{"p":"fastrouter","id":"x-ai/grok-4","n":"Grok 4","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":3,"co":15,"cr":0.75,"cw":15,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-07","rd":"2025-07-09","lu":"2025-07-09"},{"p":"fastrouter","id":"x-ai/grok-build-0.1","n":"Grok Build 0.1","f":"grok-build","d":"Fast Grok coding model tuned for agentic engineering and iterative edits","ci":1,"co":2,"ctx":256000,"out":256000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"fastrouter","id":"z-ai/glm-5.1","n":"GLM-5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.05,"co":3.5,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"fastrouter","id":"z-ai/glm-5","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.95,"co":3.15,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"fastrouter","id":"openai/gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10-01","rd":"2025-08-07","lu":"2025-08-07"},{"p":"fastrouter","id":"openai/gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"fastrouter","id":"openai/gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"fastrouter","id":"openai/gpt-realtime-1.5","n":"GPT Realtime 1.5","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":4,"co":16,"ctx":32000,"out":4096,"mi":"tai","mo":"ta","tc":1,"rd":"2025-06-01","lu":"2025-06-01"},{"p":"fastrouter","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.15,"co":0.6,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"fastrouter","id":"openai/gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"fastrouter","id":"openai/gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"fastrouter","id":"openai/gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10-01","rd":"2025-08-07","lu":"2025-08-07"},{"p":"fastrouter","id":"openai/gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-10-01","rd":"2025-08-07","lu":"2025-08-07"},{"p":"fastrouter","id":"openai/gpt-5.5-pro","n":"GPT-5.5 Pro","f":"gpt-pro","d":"Highest-accuracy GPT-5.5 tier for slower, precision-heavy reasoning and coding","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"fastrouter","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.2,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"fastrouter","id":"openai/gpt-image-2","n":"GPT Image 2","f":"gpt-image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":128000,"out":0,"mi":"ti","mo":"i","rd":"2026-04-21","lu":"2026-04-21"},{"p":"fastrouter","id":"openai/gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"fastrouter","id":"bytedance/seedance-2","n":"Seedance 2","f":"seed","d":"Video model for prompt-guided generation, editing, and motion workflows","ctx":4096,"out":0,"mi":"ti","mo":"v","rd":"2026-04-01","lu":"2026-04-01"},{"p":"fastrouter","id":"anthropic/claude-sonnet-4","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"fastrouter","id":"anthropic/claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"fastrouter","id":"anthropic/claude-opus-4.1","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"fastrouter","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"fastrouter","id":"sarvam/sarvam-105b","n":"Sarvam 105B","f":"sarvam","d":"Flagship Indian-language reasoning model for enterprise multilingual applications","ci":0.04,"co":0.16,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-09-01","lu":"2025-09-01"},{"p":"fastrouter","id":"sarvam/sarvam-30b","n":"Sarvam 30B","f":"sarvam","d":"Efficient Indian-language reasoning model for chat, coding, and multilingual work","ci":0.02,"co":0.1,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-18","lu":"2026-02-18"},{"p":"fastrouter","id":"deepseek-ai/deepseek-r1-distill-llama-70b","n":"DeepSeek R1 Distill Llama 70B","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.03,"co":0.14,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-10","rd":"2025-01-23","lu":"2025-01-23"},{"p":"fastrouter","id":"qwen/qwen3-coder","n":"Qwen3 Coder","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.3,"co":1.2,"ctx":262144,"out":66536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-23","lu":"2025-07-23"},{"p":"fastrouter","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.74,"co":3.48,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"fastrouter","id":"minimax/minimax-m2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"fastrouter","id":"minimax/minimax-m2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"fastrouter","id":"leonardo-ai/lucid-origin","n":"Lucid Origin","f":"lucid","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":4096,"out":0,"mi":"ti","mo":"i","rd":"2025-06-01","lu":"2025-06-01"},{"p":"fastrouter","id":"leonardo-ai/lucid-realism","n":"Lucid Realism","f":"lucid","d":"Image model for prompt-driven generation, editing, and visual design workflows","ctx":4096,"out":0,"mi":"ti","mo":"i","rd":"2025-06-01","lu":"2025-06-01"},{"p":"abliteration-ai","id":"abliterated-model","n":"Abliterated Model","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":3,"co":3,"ctx":150000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-01-06","lu":"2026-01-06"},{"p":"xpersona","id":"xpersona-gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":3,"co":18,"cr":0.3,"ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-12-30","rd":"2026-05-29","lu":"2026-05-29"},{"p":"xpersona","id":"xpersona-frieren-coder","n":"Xpersona Frieren 1","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":1.5,"co":6,"cr":0.15,"ctx":1000000,"out":384000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-12-30","rd":"2026-05-01","lu":"2026-05-25"},{"p":"azure-cognitive-services","id":"claude-opus-4-5","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-08-01"},{"p":"azure-cognitive-services","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-11-18","lu":"2025-11-18"},{"p":"azure-cognitive-services","id":"gpt-5.4-nano","n":"GPT-5.4 Nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"azure-cognitive-services","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-31","rd":"2026-05-28","lu":"2026-05-28"},{"p":"azure-cognitive-services","id":"claude-opus-4-1","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-18","lu":"2025-11-18"},{"p":"azure-cognitive-services","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.6,"co":3,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"azure-cognitive-services","id":"claude-haiku-4-5","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-31","rd":"2025-11-18","lu":"2025-11-18"},{"p":"azure-cognitive-services","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"azure-cognitive-services","id":"gpt-5.4-mini","n":"GPT-5.4 Mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"azure-cognitive-services","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"azure-cognitive-services","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"azure-cognitive-services","id":"gpt-5.4-pro","n":"GPT-5.4 Pro","f":"gpt-pro","d":"More exact GPT-5.4 tier for demanding professional reasoning and agent tasks","ci":30,"co":180,"ctx":1050000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"azure-cognitive-services","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-24","lu":"2026-04-24"},{"p":"azure-cognitive-services","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.125,"ctx":272000,"out":128000,"mi":"tia","mo":"tia","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure-cognitive-services","id":"meta-llama-3-70b-instruct","n":"Meta-Llama-3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.68,"co":3.54,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2023-12","rd":"2024-04-18","lu":"2024-04-18"},{"p":"azure-cognitive-services","id":"phi-3-mini-4k-instruct","n":"Phi-3-mini-instruct (4k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.13,"co":0.52,"ctx":4096,"out":1024,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure-cognitive-services","id":"phi-3-mini-128k-instruct","n":"Phi-3-mini-instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.13,"co":0.52,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure-cognitive-services","id":"phi-4-mini-reasoning","n":"Phi-4-mini-reasoning","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.075,"co":0.3,"ctx":128000,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure-cognitive-services","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-01-14","lu":"2026-01-14"},{"p":"azure-cognitive-services","id":"llama-4-maverick-17b-128e-instruct-fp8","n":"Llama 4 Maverick 17B 128E Instruct FP8","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.25,"co":1,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"azure-cognitive-services","id":"deepseek-v3.2","n":"DeepSeek-V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.58,"co":1.68,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"azure-cognitive-services","id":"gpt-5-codex","n":"GPT-5-Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-09-15","lu":"2025-09-15"},{"p":"azure-cognitive-services","id":"model-router","n":"Model Router","f":"model-router","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.14,"co":0,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2025-05-19","lu":"2025-11-18"},{"p":"azure-cognitive-services","id":"gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Small omni GPT for cheap multimodal assistance and production-scale traffic","ci":0.15,"co":0.6,"cr":0.075,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-07-18","lu":"2024-07-18"},{"p":"azure-cognitive-services","id":"phi-4-multimodal","n":"Phi-4-multimodal","f":"phi","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.08,"co":0.32,"ctx":128000,"out":4096,"mi":"tia","mo":"t","ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure-cognitive-services","id":"phi-4-reasoning-plus","n":"Phi-4-reasoning-plus","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.125,"co":0.5,"ctx":32000,"out":4096,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure-cognitive-services","id":"codestral-2501","n":"Codestral 25.01","f":"codestral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.9,"ctx":256000,"out":256000,"mi":"t","mo":"t","tc":1,"kn":"2024-03","rd":"2025-01-01","lu":"2025-01-01"},{"p":"azure-cognitive-services","id":"cohere-embed-v3-english","n":"Embed v3 English","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":512,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2023-11-07","lu":"2023-11-07"},{"p":"azure-cognitive-services","id":"phi-3-medium-4k-instruct","n":"Phi-3-medium-instruct (4k)","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.17,"co":0.68,"ctx":4096,"out":1024,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure-cognitive-services","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"cr":0.01,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure-cognitive-services","id":"gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-12","rd":"2023-11-06","lu":"2024-04-09"},{"p":"azure-cognitive-services","id":"gpt-4.1-mini","n":"GPT-4.1 mini","f":"gpt-mini","d":"Affordable GPT-4.1 lane for fast coding help and structured extraction","ci":0.4,"co":1.6,"cr":0.1,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"azure-cognitive-services","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"cr":0.03,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure-cognitive-services","id":"gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"azure-cognitive-services","id":"ministral-3b","n":"Ministral 3B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.04,"co":0.04,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-03","rd":"2024-10-22","lu":"2024-10-22"},{"p":"azure-cognitive-services","id":"llama-3.2-90b-vision-instruct","n":"Llama-3.2-90B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":2.04,"co":2.04,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"azure-cognitive-services","id":"deepseek-v3.1","n":"DeepSeek-V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.56,"co":1.68,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-08-21","lu":"2025-08-21"},{"p":"azure-cognitive-services","id":"o1","n":"o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":15,"co":60,"cr":7.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-12-05","lu":"2024-12-05"},{"p":"azure-cognitive-services","id":"gpt-4.1-nano","n":"GPT-4.1 nano","f":"gpt-nano","d":"Tiny GPT-4.1 option for classification, routing, and very high-volume tasks","ci":0.1,"co":0.4,"cr":0.025,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"azure-cognitive-services","id":"grok-4-fast-reasoning","n":"Grok 4 Fast (Reasoning)","f":"grok","d":"Fast Grok model for responsive chat, reasoning, and tool-assisted work","ci":0.2,"co":0.5,"cr":0.05,"ctx":2000000,"out":30000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-07","rd":"2025-09-19","lu":"2025-09-19"},{"p":"azure-cognitive-services","id":"phi-4","n":"Phi-4","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.125,"co":0.5,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure-cognitive-services","id":"gpt-5.1-codex-mini","n":"GPT-5.1 Codex Mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure-cognitive-services","id":"meta-llama-3.1-8b-instruct","n":"Meta-Llama-3.1-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.3,"co":0.61,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"azure-cognitive-services","id":"gpt-3.5-turbo-0301","n":"GPT-3.5 Turbo 0301","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":4096,"out":4096,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-03-01","lu":"2023-03-01"},{"p":"azure-cognitive-services","id":"deepseek-v3.2-speciale","n":"DeepSeek-V3.2-Speciale","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.58,"co":1.68,"ctx":128000,"out":128000,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"azure-cognitive-services","id":"text-embedding-3-small","n":"text-embedding-3-small","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.02,"co":0,"ctx":8191,"out":1536,"mi":"t","mo":"t","rd":"2024-01-25","lu":"2024-01-25"},{"p":"azure-cognitive-services","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-24","lu":"2026-02-24"},{"p":"azure-cognitive-services","id":"mistral-large-2411","n":"Mistral Large 24.11","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2,"co":6,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2024-09","rd":"2024-11-01","lu":"2024-11-01"},{"p":"azure-cognitive-services","id":"gpt-4-turbo-vision","n":"GPT-4 Turbo Vision","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","tc":1,"kn":"2023-11","rd":"2023-11-06","lu":"2024-04-09"},{"p":"azure-cognitive-services","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"azure-cognitive-services","id":"o3-mini","n":"o3-mini","f":"o-mini","d":"Smaller o-series reasoner for economical coding, math, and planning tasks","ci":1.1,"co":4.4,"cr":0.55,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2024-12-20","lu":"2025-01-29"},{"p":"azure-cognitive-services","id":"text-embedding-ada-002","n":"text-embedding-ada-002","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":8192,"out":1536,"mi":"t","mo":"t","rd":"2022-12-15","lu":"2022-12-15"},{"p":"azure-cognitive-services","id":"meta-llama-3.1-70b-instruct","n":"Meta-Llama-3.1-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.68,"co":3.54,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"azure-cognitive-services","id":"mistral-medium-2505","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":128000,"out":128000,"mi":"ti","mo":"t","tc":1,"kn":"2025-05","rd":"2025-05-07","lu":"2025-05-07"},{"p":"azure-cognitive-services","id":"phi-4-reasoning","n":"Phi-4-reasoning","f":"phi","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.125,"co":0.5,"ctx":32000,"out":4096,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure-cognitive-services","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"tia","mo":"tia","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure-cognitive-services","id":"gpt-3.5-turbo-0613","n":"GPT-3.5 Turbo 0613","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":4,"ctx":16384,"out":16384,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-06-13","lu":"2023-06-13"},{"p":"azure-cognitive-services","id":"cohere-embed-v3-multilingual","n":"Embed v3 Multilingual","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.1,"co":0,"ctx":512,"out":1024,"mi":"t","mo":"t","ow":1,"rd":"2023-11-07","lu":"2023-11-07"},{"p":"azure-cognitive-services","id":"gpt-3.5-turbo-0125","n":"GPT-3.5 Turbo 0125","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"ctx":16384,"out":16384,"mi":"t","mo":"t","kn":"2021-08","rd":"2024-01-25","lu":"2024-01-25"},{"p":"azure-cognitive-services","id":"phi-3-small-8k-instruct","n":"Phi-3-small-instruct (8k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.15,"co":0.6,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure-cognitive-services","id":"deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":1.35,"co":5.4,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"ow":1,"kn":"2024-07","rd":"2025-01-20","lu":"2025-01-20"},{"p":"azure-cognitive-services","id":"kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"cr":0.15,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-06","lu":"2025-12-02"},{"p":"azure-cognitive-services","id":"gpt-5.1-chat","n":"GPT-5.1 Chat","f":"gpt-codex","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":16384,"mi":"tia","mo":"tia","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-14","lu":"2025-11-14"},{"p":"azure-cognitive-services","id":"llama-3.3-70b-instruct","n":"Llama-3.3-70B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.71,"co":0.71,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-12-06","lu":"2024-12-06"},{"p":"azure-cognitive-services","id":"llama-4-scout-17b-16e-instruct","n":"Llama 4 Scout 17B 16E Instruct","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.2,"co":0.78,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-05","lu":"2025-04-05"},{"p":"azure-cognitive-services","id":"gpt-3.5-turbo-1106","n":"GPT-3.5 Turbo 1106","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":2,"ctx":16384,"out":16384,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-11-06","lu":"2023-11-06"},{"p":"azure-cognitive-services","id":"phi-4-mini","n":"Phi-4-mini","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.075,"co":0.3,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-10","rd":"2024-12-11","lu":"2024-12-11"},{"p":"azure-cognitive-services","id":"cohere-command-r-plus-08-2024","n":"Command R+","f":"command-r","d":"Cohere's RAG workhorse for long-context enterprise search and tool use","ci":2.5,"co":10,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"azure-cognitive-services","id":"meta-llama-3.1-405b-instruct","n":"Meta-Llama-3.1-405B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":5.33,"co":16,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-07-23","lu":"2024-07-23"},{"p":"azure-cognitive-services","id":"gpt-4-32k","n":"GPT-4 32K","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":60,"co":120,"ctx":32768,"out":32768,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-03-14","lu":"2023-03-14"},{"p":"azure-cognitive-services","id":"phi-3-medium-128k-instruct","n":"Phi-3-medium-instruct (128k)","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.17,"co":0.68,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure-cognitive-services","id":"o4-mini","n":"o4-mini","f":"o-mini","d":"Fast o-series model for compact reasoning, coding, and tool use","ci":1.1,"co":4.4,"cr":0.275,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"azure-cognitive-services","id":"gpt-4","n":"GPT-4","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":60,"co":120,"ctx":8192,"out":8192,"mi":"t","mo":"t","tc":1,"kn":"2023-11","rd":"2023-03-14","lu":"2023-03-14"},{"p":"azure-cognitive-services","id":"gpt-4o","n":"GPT-4o","f":"gpt","d":"Omni-era GPT for multimodal chat, practical coding, and general assistants","ci":2.5,"co":10,"cr":1.25,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"kn":"2023-09","rd":"2024-05-13","lu":"2024-08-06"},{"p":"azure-cognitive-services","id":"cohere-command-r-08-2024","n":"Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.15,"co":0.6,"ctx":128000,"out":4000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-06-01","rd":"2024-08-30","lu":"2024-08-30"},{"p":"azure-cognitive-services","id":"gpt-5-pro","n":"GPT-5 Pro","f":"gpt-pro","d":"Higher-accuracy GPT-5 tier for tough analysis, coding reviews, and planning","ci":15,"co":120,"ctx":400000,"out":272000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-10-06","lu":"2025-10-06"},{"p":"azure-cognitive-services","id":"llama-3.2-11b-vision-instruct","n":"Llama-3.2-11B-Vision-Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.37,"co":0.37,"ctx":128000,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2024-09-25","lu":"2024-09-25"},{"p":"azure-cognitive-services","id":"cohere-command-a","n":"Command A","f":"command-a","d":"Cohere command model for multilingual enterprise agents, tools, and chat","ci":2.5,"co":10,"ctx":256000,"out":8000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-06-01","rd":"2025-03-13","lu":"2025-03-13"},{"p":"azure-cognitive-services","id":"gpt-5-chat","n":"GPT-5 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.25,"co":10,"cr":0.13,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"kn":"2024-10-24","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure-cognitive-services","id":"gpt-5","n":"GPT-5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.13,"ctx":272000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"azure-cognitive-services","id":"cohere-embed-v-4-0","n":"Embed v4","f":"cohere-embed","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.12,"co":0,"ctx":128000,"out":1536,"mi":"ti","mo":"t","ow":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"azure-cognitive-services","id":"mistral-nemo","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.15,"co":0.15,"ctx":128000,"out":128000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2024-07-18","lu":"2024-07-18"},{"p":"azure-cognitive-services","id":"phi-3.5-mini-instruct","n":"Phi-3.5-mini-instruct","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.13,"co":0.52,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"azure-cognitive-services","id":"o1-mini","n":"o1-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"cr":0.55,"ctx":128000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2023-09","rd":"2024-09-12","lu":"2024-09-12"},{"p":"azure-cognitive-services","id":"text-embedding-3-large","n":"text-embedding-3-large","f":"text-embedding","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.13,"co":0,"ctx":8191,"out":3072,"mi":"t","mo":"t","rd":"2024-01-25","lu":"2024-01-25"},{"p":"azure-cognitive-services","id":"mistral-small-2503","n":"Mistral Small 3.1","f":"mistral-small","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.1,"co":0.3,"ctx":128000,"out":32768,"mi":"ti","mo":"t","tc":1,"kn":"2024-09","rd":"2025-03-01","lu":"2025-03-01"},{"p":"azure-cognitive-services","id":"meta-llama-3-8b-instruct","n":"Meta-Llama-3-8B-Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.3,"co":0.61,"ctx":8192,"out":2048,"mi":"t","mo":"t","ow":1,"kn":"2023-12","rd":"2024-04-18","lu":"2024-04-18"},{"p":"azure-cognitive-services","id":"phi-3-small-128k-instruct","n":"Phi-3-small-instruct (128k)","f":"phi","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.15,"co":0.6,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-04-23","lu":"2024-04-23"},{"p":"azure-cognitive-services","id":"deepseek-v3-0324","n":"DeepSeek-V3-0324","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":1.14,"co":4.56,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-07","rd":"2025-03-24","lu":"2025-03-24"},{"p":"azure-cognitive-services","id":"o3","n":"o3","f":"o","d":"Deliberate o-series reasoner for hard math, coding, and multi-step analysis","ci":2,"co":8,"cr":0.5,"ctx":200000,"out":100000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05","rd":"2025-04-16","lu":"2025-04-16"},{"p":"azure-cognitive-services","id":"gpt-5.2-chat","n":"GPT-5.2 Chat","f":"gpt-codex","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":1.75,"co":14,"cr":0.175,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"azure-cognitive-services","id":"deepseek-r1-0528","n":"DeepSeek-R1-0528","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":1.35,"co":5.4,"ctx":163840,"out":163840,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-05-28","lu":"2025-05-28"},{"p":"azure-cognitive-services","id":"gpt-3.5-turbo-instruct","n":"GPT-3.5 Turbo Instruct","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5,"co":2,"ctx":4096,"out":4096,"mi":"t","mo":"t","kn":"2021-08","rd":"2023-09-21","lu":"2023-09-21"},{"p":"azure-cognitive-services","id":"phi-3.5-moe-instruct","n":"Phi-3.5-MoE-instruct","f":"phi","d":"Open-weight instruction model for adaptable chat and self-hosted production workloads","ci":0.16,"co":0.64,"ctx":128000,"out":4096,"mi":"t","mo":"t","ow":1,"kn":"2023-10","rd":"2024-08-20","lu":"2024-08-20"},{"p":"azure-cognitive-services","id":"codex-mini","n":"Codex Mini","f":"gpt-codex-mini","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.5,"co":6,"cr":0.375,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-04","rd":"2025-05-16","lu":"2025-05-16"},{"p":"baseten","id":"moonshotai/Kimi-K2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"baseten","id":"moonshotai/Kimi-K2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.12,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-01-30","lu":"2026-02-12"},{"p":"baseten","id":"moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.16,"ctx":262000,"out":262000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"baseten","id":"openai/gpt-oss-120b","n":"OpenAI GPT 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.1,"co":0.5,"ctx":128072,"out":128072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-08","rd":"2025-08-05","lu":"2025-08-05"},{"p":"baseten","id":"nvidia/Nemotron-120B-A12B","n":"Nemotron Super","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.3,"co":0.75,"cr":0.06,"ctx":202800,"out":202800,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-02","rd":"2026-03-11","lu":"2026-03-11"},{"p":"baseten","id":"nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B","n":"Nemotron Ultra","f":"nemotron","d":"Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy","ci":0.6,"co":2.4,"cr":0.12,"ctx":202800,"out":202800,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-04","lu":"2026-06-04"},{"p":"baseten","id":"zai-org/GLM-5","n":"GLM 5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.95,"co":3.15,"cr":0.2,"ctx":202800,"out":202800,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"baseten","id":"zai-org/GLM-4.7","n":"GLM 4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.12,"ctx":200000,"out":200000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"baseten","id":"zai-org/GLM-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.3,"ctx":202720,"out":202720,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"baseten","id":"zai-org/GLM-5.1","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.3,"co":4.3,"cr":0.26,"ctx":202800,"out":202800,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"baseten","id":"deepseek-ai/DeepSeek-V3.1","n":"DeepSeek V3.1","f":"deepseek","d":"Legacy model retained for compatibility with older integrations","ci":0.5,"co":1.5,"ctx":164000,"out":131000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-25","lu":"2025-08-25"},{"p":"baseten","id":"deepseek-ai/DeepSeek-V4-Pro","n":"Deepseek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":1.74,"co":3.48,"cr":0.145,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"baseten","id":"MiniMaxAI/MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"Legacy model retained for compatibility with older integrations","ci":0.3,"co":1.2,"ctx":204000,"out":204000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2026-01","rd":"2026-02-12","lu":"2026-02-12"},{"p":"atomic-chat","id":"gemma-4-E4B-it-IQ4_XS","n":"Gemma 4 E4B Instruct (IQ4_XS)","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"atomic-chat","id":"Meta-Llama-3_1-8B-Instruct-GGUF","n":"Meta Llama 3.1 8B Instruct (GGUF)","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0,"co":0,"ctx":131072,"out":4096,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2024-07-23","lu":"2024-07-23"},{"p":"atomic-chat","id":"Qwen3_5-9B-MLX-4bit","n":"Qwen 3.5 9B (MLX 4-bit)","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"ctx":32768,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-03-05","lu":"2026-04-04"},{"p":"atomic-chat","id":"gemma-4-E4B-it-MLX-4bit","n":"Gemma 4 E4B Instruct (MLX 4-bit)","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0,"co":0,"ctx":32768,"out":8192,"mi":"t","mo":"t","ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"atomic-chat","id":"Qwen3_5-9B-Q4_K_M","n":"Qwen 3.5 9B (Q4_K_M)","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"ctx":32768,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-03-05","lu":"2026-04-04"},{"p":"routing-run","id":"route/deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.4928,"co":0.7392,"cr":0.0028,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"routing-run","id":"route/deepseek-v4-flash-6bit","n":"DeepSeek V4 Flash 6bit","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.4928,"co":0.7392,"cr":0.0028,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"routing-run","id":"route/minimax-m2.7-highspeed","n":"MiniMax M2.7 Highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.33,"co":1.32,"cr":0.06,"cw":0.375,"ctx":100000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"routing-run","id":"route/glm-5.1-6bit","n":"GLM 5.1 6bit","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1,"co":3,"cr":0.26,"cw":0,"ctx":202752,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"routing-run","id":"route/minimax-m2.5","n":"MiniMax M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.193,"co":1.238,"cr":0.03,"cw":0.375,"ctx":100000,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"routing-run","id":"route/mistral-small-2503","n":"Mistral Small 2503","f":"mistral-small","d":"Fast Mistral production model for chat, extraction, and cost-sensitive agents","ci":0.15,"co":0.6,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","tc":1,"ow":1,"kn":"2025-06","rd":"2026-03-16","lu":"2026-03-16"},{"p":"routing-run","id":"route/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.1,"co":0.3,"ctx":131072,"out":65536,"mi":"tiv","mo":"t","tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"routing-run","id":"route/step-3.5-flash-2603","n":"Step 3.5 Flash 2603","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"cr":0.02,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-02","lu":"2026-04-02"},{"p":"routing-run","id":"route/glm-5.1","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1,"co":3,"cr":0.26,"cw":0,"ctx":202752,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"routing-run","id":"route/deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.4928,"co":0.7392,"cr":0.003625,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"routing-run","id":"route/mistral-large-3","n":"Mistral Large 3","f":"mistral-large","d":"Mistral's largest general model for enterprise agents, coding, and multilingual reasoning","ci":0.5,"co":1.5,"ctx":128000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-11","rd":"2024-11-01","lu":"2025-12-02"},{"p":"routing-run","id":"route/deepseek-v4-pro-6bit","n":"DeepSeek V4 Pro 6bit","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.4928,"co":0.7392,"cr":0.003625,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"routing-run","id":"route/mistral-medium-2505","n":"Mistral Medium 2505","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","tc":1,"kn":"2025-05","rd":"2025-05-07","lu":"2025-05-07"},{"p":"routing-run","id":"route/minimax-m2.7","n":"MiniMax M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.33,"co":1.32,"cr":0.06,"cw":0.375,"ctx":100000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"routing-run","id":"route/qwen3.6-27b-202k","n":"Qwen3.6 27B 202K","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.1,"co":3.3,"ctx":202000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"routing-run","id":"route/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.462,"co":2.42,"cr":0.1,"ctx":131072,"out":32768,"mi":"tiv","mo":"t","tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"routing-run","id":"route/minimax-m2.5-highspeed","n":"MiniMax M2.5 Highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.193,"co":1.238,"cr":0.06,"cw":0.375,"ctx":100000,"out":131072,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"routing-run","id":"route/kimi-k2.6-6bit","n":"Kimi K2.6 6bit","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.462,"co":2.42,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"routing-run","id":"route/mimo-v2.5","n":"MiMo V2.5","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.45,"co":1.35,"cr":0.2,"ctx":1000000,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"routing-run","id":"route/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.462,"co":2.42,"cr":0.16,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"routing-run","id":"route/qwen3.6-27b","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":1.1,"co":3.3,"ctx":202000,"out":32768,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"routing-run","id":"route/mimo-v2.5-pro-6bit","n":"MiMo V2.5 Pro 6bit","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.45,"co":1.35,"cr":0.2,"ctx":1000000,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"routing-run","id":"route/stepfun-3.5-flash","n":"StepFun 3.5 Flash","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.096,"co":0.288,"cr":0.019,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-02-13"},{"p":"routing-run","id":"route/mimo-v2.5-pro","n":"MiMo V2.5 Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.45,"co":1.35,"cr":0.2,"ctx":1000000,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"routing-run","id":"route/step-3.5-flash","n":"Step 3.5 Flash","d":"StepFun flash lane for quick multimodal reasoning and coding assistance","ci":0.096,"co":0.288,"cr":0.019,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01-29","lu":"2026-02-13"},{"p":"routing-run","id":"route/deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.4928,"co":0.7392,"ctx":163840,"out":163840,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-07","rd":"2025-12-01","lu":"2025-12-01"},{"p":"aihubmix","id":"coding-minimax-m2.7","n":"Coding MiniMax M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.2,"co":0.2,"ctx":204800,"out":128100,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"aihubmix","id":"alicloud-glm-5.1","n":"GLM-5.1 (Alibaba Cloud)","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.84,"co":3.38,"cr":0.169,"cw":1.05625,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"aihubmix","id":"claude-sonnet-4-6-think","n":"Claude Sonnet 4.6 Thinking","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"aihubmix","id":"gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"cw":1,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"aihubmix","id":"alicloud-deepseek-v4-flash","n":"DeepSeek V4 Flash (Alibaba Cloud)","f":"deepseek-flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"aihubmix","id":"xiaomi-mimo-v2.5-pro","n":"Xiaomi MiMo-V2.5-Pro","f":"mimo-v2.5-pro","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":1.1,"co":3.3,"cr":0.22,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-13"},{"p":"aihubmix","id":"doubao-seed-2-0-code-preview","n":"Doubao Seed 2.0 Code Preview","f":"seed","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.48,"co":2.41,"cr":0.09644,"ctx":256000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"aihubmix","id":"coding-xiaomi-mimo-v2.5-pro","n":"Coding Xiaomi MiMo-V2.5-Pro","f":"mimo-v2.5-pro","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0.2,"co":0.6,"cr":0.04,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-13"},{"p":"aihubmix","id":"doubao-seed-2-0-pro","n":"Doubao Seed 2.0 Pro","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.48,"co":2.41,"cr":0.09644,"ctx":256000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-02-14","lu":"2026-02-14"},{"p":"aihubmix","id":"deep-deepseek-v4-flash","n":"DeepSeek V4 Flash (DeepSeek)","f":"deepseek-flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.154,"co":0.308,"cr":0.0308,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"aihubmix","id":"qwen3.7-plus","n":"Qwen3.7 Plus","f":"qwen","d":"Multimodal Qwen workhorse for long-context agents, visual inputs, and coding","ci":0.282,"co":1.128,"cr":0.0564,"cw":0.3525,"ctx":991000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-06-02","lu":"2026-06-02"},{"p":"aihubmix","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-05"},{"p":"aihubmix","id":"grok-4.3","n":"Grok 4.3","f":"grok","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"aihubmix","id":"qwen3.7-max","n":"Qwen3.7 Max","f":"qwen","d":"Qwen frontier model tuned for agent frameworks, coding assistants, and long tasks","ci":1.69,"co":5.07,"cr":0.169,"cw":2.1125,"ctx":991000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"aihubmix","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-03-20","lu":"2025-06-05"},{"p":"aihubmix","id":"xiaomi-mimo-v2.5-free","n":"Xiaomi MiMo-V2.5 (free)","f":"mimo-v2.5","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-13"},{"p":"aihubmix","id":"claude-opus-4-7-think","n":"Claude Opus 4.7 Thinking","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"aihubmix","id":"coding-glm-5.1-free","n":"Coding GLM 5.1 (free)","f":"glm-free","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-11","lu":"2026-04-11"},{"p":"aihubmix","id":"xiaomi-mimo-v2.5-pro-free","n":"Xiaomi MiMo-V2.5-Pro (free)","f":"mimo-v2.5-pro","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-13"},{"p":"aihubmix","id":"xiaomi-mimo-v2.5","n":"Xiaomi MiMo-V2.5","f":"mimo-v2.5","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.44,"co":2.2,"cr":0.088,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-13"},{"p":"aihubmix","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"aihubmix","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.1268,"co":3.9438,"cr":0.2817,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"aihubmix","id":"qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.17,"co":1.01,"cr":0.0169,"cw":0.21125,"ctx":991000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-04-02","lu":"2026-04-02"},{"p":"aihubmix","id":"coding-xiaomi-mimo-v2.5","n":"Coding Xiaomi MiMo-V2.5","f":"mimo-v2.5","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0.08,"co":0.4,"cr":0.016,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-05-13"},{"p":"aihubmix","id":"gpt-5.1-codex","n":"GPT-5.1 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"aihubmix","id":"claude-opus-4-8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"aihubmix","id":"gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"aihubmix","id":"doubao-seed-2-0-mini-260428","n":"Doubao Seed 2.0 Mini 260428","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.03,"co":0.28,"cr":0.00564,"ctx":256000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"aihubmix","id":"doubao-seed-2-0-lite-260428","n":"Doubao Seed 2.0 Lite 260428","f":"seed","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.08,"co":0.51,"cr":0.01692,"ctx":256000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"aihubmix","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"aihubmix","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"aihubmix","id":"minimax-m2.7","n":"MiniMax M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"aihubmix","id":"claude-opus-4-6-think","n":"Claude Opus 4.6 Thinking","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"aihubmix","id":"gpt-5.1-codex-mini","n":"GPT-5.1 Codex mini","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"aihubmix","id":"kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.6,"co":3,"cr":0.1,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-01","lu":"2026-01"},{"p":"aihubmix","id":"coding-glm-5.1","n":"Coding GLM 5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.06,"co":0.22,"cr":0.013,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-11","lu":"2026-04-11"},{"p":"aihubmix","id":"coding-minimax-m2.7-free","n":"Coding MiniMax M2.7 (Free)","f":"minimax-free","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"ctx":204800,"out":128100,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"aihubmix","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"aihubmix","id":"gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"aihubmix","id":"kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.95,"co":4,"cr":0.16,"ctx":262144,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"aihubmix","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"aihubmix","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"aihubmix","id":"deep-deepseek-v4-pro","n":"DeepSeek V4 Pro (DeepSeek)","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.478,"co":0.956,"cr":0.004302,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"aihubmix","id":"glm-5v-turbo","n":"GLM 5 Vision Turbo","f":"glmv","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.7042,"co":3.09848,"cr":0.169008,"ctx":200000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-05-09","lu":"2026-05-09"},{"p":"aihubmix","id":"zai-glm-5.1","n":"GLM-5.1 (Z.ai)","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.845,"co":3.38,"cr":0.183112,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"aihubmix","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"aihubmix","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"aihubmix","id":"coding-minimax-m2.7-highspeed","n":"Coding MiniMax M2.7 Highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.2,"co":0.2,"ctx":204800,"out":128100,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"aihubmix","id":"claude-opus-4-8-think","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"aihubmix","id":"qwen3.6-max-preview","n":"Qwen3.6 Max Preview","f":"qwen3.6","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":1.27,"co":7.61,"cr":0.1268,"cw":1.585,"ctx":240000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-05-09","lu":"2026-05-09"},{"p":"aihubmix","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"aihubmix","id":"qwen3.6-plus","n":"Qwen3.6 Plus","f":"qwen3.6","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.28,"co":1.69,"cr":0.0282,"cw":0.3525,"ctx":991000,"out":64000,"mi":"tiv","mo":"t","rs":1,"tc":1,"kn":"2025-04","rd":"2026-05-09","lu":"2026-05-09"},{"p":"aihubmix","id":"alicloud-deepseek-v4-pro","n":"DeepSeek V4 Pro (Alibaba Cloud)","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.69,"co":3.38,"cr":0.13,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"aihubmix","id":"gpt-5.1","n":"GPT-5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.13,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"aihubmix","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"google-vertex","id":"gemini-2.5-pro-tts","n":"Gemini 2.5 Pro TTS","f":"gemini-pro","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":1,"co":20,"ctx":32768,"out":16384,"mi":"t","mo":"a","kn":"2025-01","rd":"2025-09-30","lu":"2025-12-10"},{"p":"google-vertex","id":"claude-haiku-4-5@20251001","n":"Claude Haiku 4.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"google-vertex","id":"gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-07","lu":"2026-05-07"},{"p":"google-vertex","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"google-vertex","id":"gemini-2.5-flash-tts","n":"Gemini 2.5 Flash TTS","f":"gemini-flash","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0.5,"co":10,"ctx":32768,"out":16384,"mi":"t","mo":"a","kn":"2025-01","rd":"2025-09-30","lu":"2025-12-10"},{"p":"google-vertex","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"cw":0.383,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"google-vertex","id":"gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"google-vertex","id":"claude-opus-4@20250514","n":"Claude Opus 4","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"google-vertex","id":"claude-opus-4-1@20250805","n":"Claude Opus 4.1","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"google-vertex","id":"gemini-embedding-001","n":"Gemini Embedding 001","f":"gemini","d":"Embedding model for semantic search, retrieval, clustering, and ranking pipelines","ci":0.15,"co":0,"ctx":2048,"out":1,"mi":"t","mo":"t","kn":"2025-05","rd":"2025-05-20","lu":"2025-05-20"},{"p":"google-vertex","id":"claude-opus-4-5@20251101","n":"Claude Opus 4.5","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-01","lu":"2025-11-01"},{"p":"google-vertex","id":"claude-3-5-haiku@20241022","n":"Claude Haiku 3.5","f":"claude-haiku","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":0.8,"co":4,"cr":0.08,"cw":1,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"kn":"2024-07-31","rd":"2024-10-22","lu":"2024-10-22"},{"p":"google-vertex","id":"gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"google-vertex","id":"gemini-flash-lite-latest","n":"Gemini Flash-Lite Latest","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.1,"co":0.4,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"google-vertex","id":"claude-sonnet-4@20250514","n":"Claude Sonnet 4","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"google-vertex","id":"gemini-2.5-flash-lite","n":"Gemini 2.5 Flash-Lite","f":"gemini-flash-lite","d":"Lean Gemini 2.5 lane for cheap multimodal traffic and quick agents","ci":0.1,"co":0.4,"cr":0.01,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"google-vertex","id":"claude-opus-4-7@default","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"google-vertex","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"google-vertex","id":"claude-sonnet-4-5@20250929","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"google-vertex","id":"claude-sonnet-5@default","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"google-vertex","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"google-vertex","id":"claude-opus-4-6@default","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"google-vertex","id":"gemini-flash-latest","n":"Gemini Flash Latest","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.3,"co":2.5,"cr":0.075,"cw":0.383,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-09-25","lu":"2025-09-25"},{"p":"google-vertex","id":"claude-opus-4-8@default","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"google-vertex","id":"claude-sonnet-4-6@default","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"google-vertex","id":"gemini-3.1-flash-lite-preview","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"google-vertex","id":"moonshotai/kimi-k2-thinking-maas","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-08","rd":"2025-11-13","lu":"2025-11-13"},{"p":"google-vertex","id":"openai/gpt-oss-120b-maas","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.09,"co":0.36,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"google-vertex","id":"openai/gpt-oss-20b-maas","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.07,"co":0.25,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"google-vertex","id":"zai-org/glm-4.7-maas","n":"GLM-4.7","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":2.2,"ctx":200000,"out":128000,"mi":"tp","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-06","lu":"2026-01-06"},{"p":"google-vertex","id":"zai-org/glm-5-maas","n":"GLM-5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.1,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"google-vertex","id":"deepseek-ai/deepseek-v3.1-maas","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.6,"co":1.7,"ctx":163840,"out":32768,"mi":"tp","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-28","lu":"2025-08-28"},{"p":"google-vertex","id":"deepseek-ai/deepseek-v3.2-maas","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.56,"co":1.68,"cr":0.056,"ctx":163840,"out":65536,"mi":"tp","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-17","lu":"2026-04-04"},{"p":"google-vertex","id":"qwen/qwen3-235b-a22b-instruct-2507-maas","n":"Qwen3 235B A22B Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.22,"co":0.88,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-13","lu":"2025-08-13"},{"p":"google-vertex","id":"meta/llama-3.3-70b-instruct-maas","n":"Llama 3.3 70B Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.72,"co":0.72,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2023-12","rd":"2025-04-29","lu":"2025-04-29"},{"p":"google-vertex","id":"meta/llama-4-maverick-17b-128e-instruct-maas","n":"Llama 4 Maverick 17B 128E Instruct","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.35,"co":1.15,"ctx":524288,"out":8192,"mi":"ti","mo":"t","tc":1,"ow":1,"kn":"2024-08","rd":"2025-04-29","lu":"2025-04-29"},{"p":"nano-gpt","id":"step-3","n":"Step-3","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2499,"co":0.6494,"ctx":65536,"out":8192,"mi":"ti","mo":"t","rd":"2025-07-31","lu":"2025-07-31"},{"p":"nano-gpt","id":"qwen3.5-35b-a3b:thinking","n":"Qwen3.5 35B A3B Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.225,"co":1.8,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"glm-4.1v-thinking-flashx","n":"GLM 4.1V Thinking FlashX","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":0.3,"ctx":64000,"out":8192,"mi":"ti","mo":"t","rd":"2025-07-09","lu":"2025-07-09"},{"p":"nano-gpt","id":"ernie-x1.1-preview","n":"ERNIE X1.1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":64000,"out":8192,"mi":"tp","mo":"t","rd":"2025-09-10","lu":"2025-09-10"},{"p":"nano-gpt","id":"qwen25-vl-72b-instruct","n":"Qwen25 VL 72b","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.69989,"co":0.69989,"ctx":32000,"out":32768,"mi":"ti","mo":"t","rd":"2025-05-10","lu":"2025-05-10"},{"p":"nano-gpt","id":"gemini-2.0-pro-exp-02-05","n":"Gemini 2.0 Pro 0205","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.989,"co":7.956,"ctx":2097152,"out":8192,"mi":"ti","mo":"t","rd":"2025-02-05","lu":"2025-02-05"},{"p":"nano-gpt","id":"doubao-seed-2-0-lite-260215","n":"Doubao Seed 2.0 Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1462,"co":0.8738,"ctx":256000,"out":32000,"mi":"t","mo":"t","rd":"2026-02-14","lu":"2026-02-14"},{"p":"nano-gpt","id":"Qwen3.5-27B-Writer-V2-Derestricted","n":"Qwen3.5 27B Writer V2 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"claude-opus-4-thinking:8192","n":"Claude 4 Opus Thinking (8K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"qwen3-vl-235b-a22b-thinking","n":"Qwen3 VL 235B A22B Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":6,"ctx":32768,"out":32768,"mi":"ti","mo":"t","rs":1,"rd":"2025-08-26","lu":"2025-08-26"},{"p":"nano-gpt","id":"glm-4-air-0111","n":"GLM 4 Air 0111","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1394,"co":0.1394,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2025-01-11","lu":"2025-01-11"},{"p":"nano-gpt","id":"Qwen3.5-27B-Queen-Derestricted","n":"Qwen3.5 27B Queen Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"gemini-2.5-pro-preview-03-25","n":"Gemini 2.5 Pro Preview 0325","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":10,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-03-25","lu":"2025-03-25"},{"p":"nano-gpt","id":"brave-research","n":"Brave (Research)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":5,"co":5,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2023-03-02","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen-plus","n":"Qwen Plus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3995,"co":1.2002,"ctx":995904,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2024-01-25","lu":"2024-01-25"},{"p":"nano-gpt","id":"doubao-seed-2-0-mini-260215","n":"Doubao Seed 2.0 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.0493,"co":0.4845,"ctx":256000,"out":32000,"mi":"t","mo":"t","rd":"2026-02-14","lu":"2026-02-14"},{"p":"nano-gpt","id":"Qwen3.5-27B-BlueStar-v3-Derestricted-Lite","n":"Qwen3.5 27B BlueStar v3 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"deepseek-v3-0324","n":"DeepSeek Chat 0324","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0.25,"co":0.7,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-03-24","lu":"2025-03-24"},{"p":"nano-gpt","id":"Gemma-4-31B-Musica-v1","n":"Gemma 4 31B Musica v1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"mirothinker-1-7-deepresearch-mini","n":"MiroThinker 1.7 Deep Research Mini","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":1.25,"co":10,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"Baichuan4-Turbo","n":"Baichuan 4 Turbo","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.42,"co":2.42,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-08-19","lu":"2025-08-19"},{"p":"nano-gpt","id":"doubao-seed-1-6-flash-250615","n":"Doubao Seed 1.6 Flash","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.0374,"co":0.374,"ctx":256000,"out":16384,"mi":"t","mo":"t","rd":"2025-06-15","lu":"2025-06-15"},{"p":"nano-gpt","id":"kimi-k2-instruct-fast","n":"Kimi K2 0711 Fast","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":2,"ctx":131072,"out":16384,"mi":"tp","mo":"t","rd":"2025-07-15","lu":"2025-07-15"},{"p":"nano-gpt","id":"glm-4-plus","n":"GLM-4 Plus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":7.497,"co":7.497,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2024-08-01","lu":"2024-08-01"},{"p":"nano-gpt","id":"v0-1.0-md","n":"v0 1.0 MD","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"t","mo":"t","rd":"2025-07-04","lu":"2025-07-04"},{"p":"nano-gpt","id":"qwen3-coder-30b-a3b-instruct","n":"Qwen3 Coder 30B A3B Instruct","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":128000,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"nano-gpt","id":"gemini-2.5-flash-preview-09-2025-thinking","n":"Gemini 2.5 Flash Preview (09/2025) – Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":2.5,"ctx":1048756,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"v0-1.5-lg","n":"v0 1.5 LG","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":15,"co":75,"ctx":1000000,"out":64000,"mi":"t","mo":"t","rd":"2025-07-04","lu":"2025-07-04"},{"p":"nano-gpt","id":"qwen3.5-flash:thinking","n":"Qwen3.5 Flash Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.09,"co":0.36,"ctx":991808,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"jamba-large","n":"Jamba Large","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.989,"co":7.99,"ctx":256000,"out":4096,"mi":"t","mo":"t","rd":"2025-07-09","lu":"2025-07-09"},{"p":"nano-gpt","id":"kimi-thinking-preview","n":"Kimi Thinking Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":31.46,"co":31.46,"ctx":128000,"out":16384,"mi":"tp","mo":"t","rd":"2025-05-07","lu":"2025-05-07"},{"p":"nano-gpt","id":"gemini-2.0-flash-thinking-exp-01-21","n":"Gemini 2.0 Flash Thinking 0121","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":1.003,"ctx":1000000,"out":8192,"mi":"ti","mo":"t","rs":1,"rd":"2025-01-21","lu":"2025-01-21"},{"p":"nano-gpt","id":"claude-opus-4-1-thinking:1024","n":"Claude 4.1 Opus Thinking (1K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"Qwen3.5-27B-NaNovel-Derestricted-Lite","n":"Qwen3.5 27B NaNovel Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"doubao-seed-1-6-250615","n":"Doubao Seed 1.6","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.204,"co":0.51,"ctx":256000,"out":16384,"mi":"t","mo":"t","rd":"2025-06-15","lu":"2025-06-15"},{"p":"nano-gpt","id":"ernie-5.0-thinking-preview","n":"Ernie 5.0 Thinking Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.1,"co":2,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rs":1,"rd":"2025-11-18","lu":"2025-11-18"},{"p":"nano-gpt","id":"gemini-2.5-flash-preview-05-20:thinking","n":"Gemini 2.5 Flash 0520 Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":3.5,"ctx":1048000,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-05-20","lu":"2025-05-20"},{"p":"nano-gpt","id":"Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted","n":"Qwen3.5 27B Omega Evolution v2.2 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"azure-o1","n":"Azure o1","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":14.994,"co":59.993,"ctx":200000,"out":100000,"mi":"t","mo":"t","rd":"2024-12-17","lu":"2024-12-17"},{"p":"nano-gpt","id":"qwen3.7-plus","n":"Qwen3.7 Plus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"cr":0.04,"ctx":991808,"out":65536,"mi":"tiv","mo":"t","rd":"2026-06-01","lu":"2026-06-01"},{"p":"nano-gpt","id":"glm-4-airx","n":"GLM-4 AirX","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.006,"co":2.006,"ctx":8000,"out":4096,"mi":"t","mo":"t","rd":"2024-06-05","lu":"2024-06-05"},{"p":"nano-gpt","id":"step-r1-v-mini","n":"Step R1 V Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":11,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-04-08","lu":"2025-04-08"},{"p":"nano-gpt","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":10,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-06-05","lu":"2025-06-05"},{"p":"nano-gpt","id":"claude-haiku-4-5-20251001-thinking","n":"Claude Haiku 4.5 Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":5,"cr":0.1,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"nano-gpt","id":"jamba-mini-1.6","n":"Jamba Mini 1.6","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1989,"co":0.408,"ctx":256000,"out":4096,"mi":"t","mo":"t","rd":"2025-03-01","lu":"2025-03-01"},{"p":"nano-gpt","id":"claude-haiku-4-5-20251001","n":"Claude Haiku 4.5","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":5,"ctx":200000,"out":64000,"mi":"tip","mo":"t","tc":1,"rd":"2025-10-15","lu":"2025-10-15"},{"p":"nano-gpt","id":"claude-3-5-haiku-20241022","n":"Claude 3.5 Haiku","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.8,"co":4,"ctx":200000,"out":8192,"mi":"tip","mo":"t","tc":1,"rd":"2024-10-22","lu":"2024-10-22"},{"p":"nano-gpt","id":"claude-sonnet-4-thinking:32768","n":"Claude 4 Sonnet Thinking (32K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"hermes-low","n":"Hermes Low","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"claude-sonnet-4-thinking","n":"Claude 4 Sonnet Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-02-24","lu":"2025-02-24"},{"p":"nano-gpt","id":"qwen3.7-max","n":"Qwen3.7 Max","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":7.5,"cr":0.25,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rd":"2026-05-21","lu":"2026-05-21"},{"p":"nano-gpt","id":"yi-medium-200k","n":"Yi Medium 200k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.499,"co":2.499,"ctx":200000,"out":4096,"mi":"t","mo":"t","rd":"2024-03-01","lu":"2024-03-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-BlueStar-v2-Derestricted-Lite","n":"Qwen3.5 27B BlueStar v2 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"Gemma-4-31B-Claude-4.6-Opus-Reasoning-Distilled","n":"Gemma 4 31B Claude 4.6 Opus Reasoning Distilled","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0.306,"co":0.306,"cr":0.0306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"doubao-1-5-thinking-vision-pro-250428","n":"Doubao 1.5 Thinking Vision Pro","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.55,"co":1.43,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2025-05-15","lu":"2025-05-15"},{"p":"nano-gpt","id":"Qwen3.5-27B-BlueStar-Derestricted-Lite","n":"Qwen3.5 27B BlueStar Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"Qwen2.5-32B-EVA-v0.2","n":"Qwen 2.5 32b EVA","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.493,"co":0.493,"ctx":24576,"out":8192,"mi":"t","mo":"t","rd":"2024-09-01","lu":"2024-09-01"},{"p":"nano-gpt","id":"gemini-2.5-flash","n":"Gemini 2.5 Flash","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":2.5,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-06-05","lu":"2025-06-05"},{"p":"nano-gpt","id":"qwen-long","n":"Qwen Long 10M","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1003,"co":0.408,"ctx":10000000,"out":8192,"mi":"tp","mo":"t","rd":"2025-01-25","lu":"2025-01-25"},{"p":"nano-gpt","id":"Gemma-4-31B-DarkIdol","n":"Gemma 4 31B DarkIdol","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"step-2-16k-exp","n":"Step-2 16k Exp","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":7.004,"co":19.992,"ctx":16000,"out":8192,"mi":"t","mo":"t","rd":"2024-07-05","lu":"2024-07-05"},{"p":"nano-gpt","id":"Qwen3.5-27B-Marvin-V2-Derestricted-Lite","n":"Qwen3.5 27B Marvin V2 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"Qwen3.5-27B-NaNovel-Derestricted","n":"Qwen3.5 27B NaNovel Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"ernie-4.5-8k-preview","n":"Ernie 4.5 8k Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.66,"co":2.6,"ctx":8000,"out":16384,"mi":"t","mo":"t","rd":"2025-03-25","lu":"2025-03-25"},{"p":"nano-gpt","id":"gemini-2.0-flash-exp-image-generation","n":"Gemini Text + Image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":0.2,"co":0.8,"ctx":32767,"out":8192,"mi":"t","mo":"t","rd":"2025-02-19","lu":"2025-02-19"},{"p":"nano-gpt","id":"gemini-2.5-flash-lite-preview-09-2025","n":"Gemini 2.5 Flash Lite Preview (09/2025)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":1048756,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"deepseek-reasoner-cheaper","n":"Deepseek R1 Cheaper","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.7,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-01-20","lu":"2025-01-20"},{"p":"nano-gpt","id":"venice-uncensored","n":"Venice Uncensored","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":0.4,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-02-24","lu":"2025-02-24"},{"p":"nano-gpt","id":"gemini-2.0-flash-001","n":"Gemini 2.0 Flash","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1003,"co":0.408,"ctx":1000000,"out":8192,"mi":"ti","mo":"t","tc":1,"rd":"2024-12-11","lu":"2024-12-11"},{"p":"nano-gpt","id":"gemma-4-31B-Larkspur-v0.5","n":"Gemma 4 31B Larkspur v0.5","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"claude-opus-4-1-20250805","n":"Claude 4.1 Opus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"nano-gpt","id":"holo3-35b-a3b","n":"Holo3-35B-A3B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":1.8,"ctx":65536,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen3.5-omni-plus","n":"Qwen3.5 Omni Plus","d":"Omni-modal model for text, vision, audio, and multimodal agent tasks","ci":0,"co":0,"ctx":983616,"out":65536,"mi":"tiva","mo":"t","rd":"2026-03-30","lu":"2026-03-30"},{"p":"nano-gpt","id":"Gemma-4-31B-it","n":"Gemma 4 31B IT","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-09","lu":"2026-04-09"},{"p":"nano-gpt","id":"Qwen3.5-27B-Derestricted","n":"Qwen3.5 27B Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-03-17","lu":"2026-03-17"},{"p":"nano-gpt","id":"sonar-reasoning-pro","n":"Perplexity Reasoning Pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2.006,"co":7.9985,"ctx":127000,"out":128000,"mi":"t","mo":"t","rs":1,"rd":"2025-02-19","lu":"2025-02-19"},{"p":"nano-gpt","id":"venice-uncensored:web","n":"Venice Uncensored Web","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":0.4,"ctx":80000,"out":16384,"mi":"t","mo":"t","rd":"2024-05-01","lu":"2024-05-01"},{"p":"nano-gpt","id":"doubao-1.5-pro-32k","n":"Doubao 1.5 Pro 32k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1343,"co":0.3349,"ctx":32000,"out":8192,"mi":"t","mo":"t","rd":"2025-01-22","lu":"2025-01-22"},{"p":"nano-gpt","id":"qwen3-30b-a3b-instruct-2507","n":"Qwen3 30B A3B Instruct 2507","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":0.5,"ctx":256000,"out":32768,"mi":"t","mo":"t","rd":"2025-02-20","lu":"2025-02-20"},{"p":"nano-gpt","id":"ernie-5.1:thinking","n":"ERNIE 5.1 Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":3,"cr":0.75,"ctx":119000,"out":64000,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-10","lu":"2026-05-10"},{"p":"nano-gpt","id":"mirothinker-1-7-deepresearch","n":"MiroThinker 1.7 Deep Research","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":4,"co":25,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"hermes-high","n":"Hermes High","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"jamba-large-1.6","n":"Jamba Large 1.6","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.989,"co":7.99,"ctx":256000,"out":4096,"mi":"t","mo":"t","rd":"2025-03-12","lu":"2025-03-12"},{"p":"nano-gpt","id":"glm-4-long","n":"GLM-4 Long","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2006,"co":0.2006,"ctx":1000000,"out":4096,"mi":"t","mo":"t","rd":"2024-08-01","lu":"2024-08-01"},{"p":"nano-gpt","id":"claude-opus-4-thinking:32768","n":"Claude 4 Opus Thinking (32K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"claude-sonnet-4-thinking:8192","n":"Claude 4 Sonnet Thinking (8K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"azure-o3-mini","n":"Azure o3-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.088,"co":4.3996,"ctx":200000,"out":65536,"mi":"t","mo":"t","rd":"2025-01-31","lu":"2025-01-31"},{"p":"nano-gpt","id":"qvq-max","n":"Qwen: QvQ Max","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.4,"co":5.3,"ctx":128000,"out":8192,"mi":"ti","mo":"t","rd":"2025-03-28","lu":"2025-03-28"},{"p":"nano-gpt","id":"Qwen3.5-27B-BlueStar-v2-Derestricted","n":"Qwen3.5 27B BlueStar v2 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"Qwen3.5-27B-Vivid-Durian","n":"Qwen3.5 27B Vivid Durian","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"nano-gpt","id":"jamba-large-1.7","n":"Jamba Large 1.7","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.989,"co":7.99,"ctx":256000,"out":4096,"mi":"t","mo":"t","rd":"2025-07-09","lu":"2025-07-09"},{"p":"nano-gpt","id":"Baichuan-M2","n":"Baichuan M2 32B Medical","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":15.73,"co":15.73,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-08-19","lu":"2025-08-19"},{"p":"nano-gpt","id":"Magistral-Small-2506","n":"Magistral Small 2506","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.4,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"deepseek-r1","n":"DeepSeek R1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.7,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"nano-gpt","id":"Qwen3.5-27B-earica-Derestricted-Lite","n":"Qwen3.5 27B earica Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"doubao-seed-2-0-pro-260215","n":"Doubao Seed 2.0 Pro","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.782,"co":3.876,"ctx":256000,"out":128000,"mi":"t","mo":"t","rd":"2026-02-14","lu":"2026-02-14"},{"p":"nano-gpt","id":"step-2-mini","n":"Step-2 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2006,"co":0.408,"ctx":8000,"out":4096,"mi":"t","mo":"t","rd":"2024-07-05","lu":"2024-07-05"},{"p":"nano-gpt","id":"gemini-exp-1206","n":"Gemini 2.0 Pro 1206","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.258,"co":4.998,"ctx":2097152,"out":8192,"mi":"ti","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"claude-opus-4-5-20251101:thinking","n":"Claude 4.5 Opus Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":4.998,"co":25.007,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-11-01","lu":"2025-11-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-BlueStar-v3-Derestricted","n":"Qwen3.5 27B BlueStar v3 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"claude-opus-4-thinking","n":"Claude 4 Opus Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-07-15","lu":"2025-07-15"},{"p":"nano-gpt","id":"qwen3.5-flash","n":"Qwen3.5 Flash","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.09,"co":0.36,"ctx":991808,"out":65536,"mi":"tiv","mo":"t","rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"exa-research-pro","n":"Exa (Research Pro)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":2.5,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2025-06-04","lu":"2025-06-04"},{"p":"nano-gpt","id":"jamba-mini","n":"Jamba Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1989,"co":0.408,"ctx":256000,"out":4096,"mi":"t","mo":"t","rd":"2025-07-09","lu":"2025-07-09"},{"p":"nano-gpt","id":"claude-opus-4-1-thinking:32000","n":"Claude 4.1 Opus Thinking (32K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"KAT-Coder-Air-V1","n":"KAT Coder Air V1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.2,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-10-28","lu":"2025-10-28"},{"p":"nano-gpt","id":"exa-research","n":"Exa (Research)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":2.5,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2025-06-04","lu":"2025-06-04"},{"p":"nano-gpt","id":"deepclaude","n":"DeepClaude","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":15,"ctx":128000,"out":8192,"mi":"tp","mo":"t","rd":"2025-02-01","lu":"2025-02-01"},{"p":"nano-gpt","id":"gemini-2.5-flash-preview-09-2025","n":"Gemini 2.5 Flash Preview (09/2025)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":2.5,"ctx":1048756,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"claude-opus-4-5-20251101","n":"Claude 4.5 Opus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":4.998,"co":25.007,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-11-01","lu":"2025-11-01"},{"p":"nano-gpt","id":"GLM-4.6-Derestricted-v5","n":"GLM 4.6 Derestricted v5","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.5,"ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-12-23","lu":"2025-12-23"},{"p":"nano-gpt","id":"glm-z1-airx","n":"GLM Z1 AirX","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.7,"co":0.7,"ctx":32000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"claude-sonnet-4-thinking:1024","n":"Claude 4 Sonnet Thinking (1K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"holo3-35b-a3b:thinking","n":"Holo3-35B-A3B Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":1.8,"ctx":65536,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"owl","n":"OWL","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.3,"ctx":1048756,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"gemini-2.0-pro-reasoner","n":"Gemini 2.0 Pro Reasoner","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.292,"co":4.998,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-02-05","lu":"2025-02-05"},{"p":"nano-gpt","id":"Gemma-4-31B-Cognitive-Unshackled","n":"Gemma 4 31B Cognitive Unshackled","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-earica-Derestricted","n":"Qwen3.5 27B earica Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"command-a-plus-05-2026","n":"Cohere Command A+ (05/2026)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":10,"ctx":128000,"out":64000,"mi":"ti","mo":"t","rs":1,"rd":"2026-05-22","lu":"2026-05-22"},{"p":"nano-gpt","id":"auto-model-premium","n":"Auto model (Premium)","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":9.996,"co":19.992,"ctx":1000000,"out":1000000,"mi":"t","mo":"t","rd":"2024-06-01","lu":"2024-06-01"},{"p":"nano-gpt","id":"learnlm-1.5-pro-experimental","n":"Gemini LearnLM Experimental","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3.502,"co":10.506,"ctx":32767,"out":8192,"mi":"t","mo":"t","rd":"2024-05-14","lu":"2024-05-14"},{"p":"nano-gpt","id":"deepseek-r1-sambanova","n":"DeepSeek R1 Fast","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":4.998,"co":6.987,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2025-02-20","lu":"2025-02-20"},{"p":"nano-gpt","id":"claw-medium","n":"Claw Medium","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"claude-opus-4-20250514","n":"Claude 4 Opus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","tc":1,"rd":"2025-05-14","lu":"2025-05-14"},{"p":"nano-gpt","id":"yi-large","n":"Yi Large","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3.196,"co":3.196,"ctx":32000,"out":4096,"mi":"t","mo":"t","rd":"2024-05-13","lu":"2024-05-13"},{"p":"nano-gpt","id":"qwen3-max-2026-01-23","n":"Qwen3 Max 2026-01-23","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.2002,"co":6.001,"ctx":256000,"out":32768,"mi":"t","mo":"t","rd":"2026-01-26","lu":"2026-01-26"},{"p":"nano-gpt","id":"phi-4-mini-instruct","n":"Phi 4 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.17,"co":0.68,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"ernie-x1-turbo-32k","n":"Ernie X1 Turbo 32k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.165,"co":0.66,"ctx":32000,"out":16384,"mi":"tip","mo":"t","rd":"2025-05-08","lu":"2025-05-08"},{"p":"nano-gpt","id":"claw-low","n":"Claw Low","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"gemini-3-pro-image-preview","n":"Gemini 3 Pro Image","d":"Image model for prompt-driven generation, editing, and visual design workflows","ci":2,"co":12,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rd":"2025-11-18","lu":"2025-11-18"},{"p":"nano-gpt","id":"gemma-4-31B-Garnet","n":"Gemma 4 31B Garnet","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"doubao-1.5-vision-pro-32k","n":"Doubao 1.5 Vision Pro 32k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.459,"co":1.377,"ctx":32000,"out":8192,"mi":"ti","mo":"t","rd":"2025-01-22","lu":"2025-01-22"},{"p":"nano-gpt","id":"auto-model-standard","n":"Auto model (Standard)","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":9.996,"co":19.992,"ctx":1000000,"out":1000000,"mi":"t","mo":"t","rd":"2024-06-01","lu":"2024-06-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-Marvin-DPO-V2-Derestricted-Lite","n":"Qwen3.5 27B Marvin DPO V2 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"glm-4","n":"GLM-4","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":14.994,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2024-01-16","lu":"2024-01-16"},{"p":"nano-gpt","id":"Qwen3.5-27B-Writer-Derestricted-Lite","n":"Qwen3.5 27B Writer Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"qwen-3.6-plus","n":"Qwen 3.6 Plus","f":"qwen3.6","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.45,"co":2.7,"ctx":991800,"out":65536,"mi":"tiv","mo":"t","rd":"2026-04-02","lu":"2026-04-02"},{"p":"nano-gpt","id":"brave-pro","n":"Brave (Pro)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":5,"co":5,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2023-03-02","lu":"2024-01-01"},{"p":"nano-gpt","id":"deepseek-chat-cheaper","n":"DeepSeek V3/Chat Cheaper","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0.25,"co":0.7,"ctx":128000,"out":8192,"mi":"tp","mo":"t","tc":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"Gemma-4-31B-Queen","n":"Gemma 4 31B Queen","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"Gemma-4-31B-Gemopus","n":"Gemma 4 31B Gemopus","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"mistral-code-latest","n":"Mistral Code Latest","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":0.9,"ctx":256000,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-06-02","lu":"2026-06-02"},{"p":"nano-gpt","id":"gemini-2.5-flash-lite","n":"Gemini 2.5 Flash Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-06-17","lu":"2025-06-17"},{"p":"nano-gpt","id":"jamba-mini-1.7","n":"Jamba Mini 1.7","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1989,"co":0.408,"ctx":256000,"out":4096,"mi":"t","mo":"t","rd":"2025-07-09","lu":"2025-07-09"},{"p":"nano-gpt","id":"universal-summarizer","n":"Universal Summarizer","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":30,"co":30,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2023-05-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-Anko","n":"Qwen3.5 27B Anko","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"claude-sonnet-4-20250514","n":"Claude 4 Sonnet","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":200000,"out":64000,"mi":"tip","mo":"t","tc":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"nano-gpt","id":"ernie-x1-32k-preview","n":"Ernie X1 32k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.33,"co":1.32,"ctx":32000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-03","lu":"2025-04-03"},{"p":"nano-gpt","id":"azure-gpt-4-turbo","n":"Azure gpt-4-turbo","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":9.996,"co":30.005,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2023-11-06","lu":"2024-01-01"},{"p":"nano-gpt","id":"Meta-Llama-3-1-8B-Instruct-FP8","n":"Llama 3.1 8B (decentralized)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.02,"co":0.03,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2024-07-23","lu":"2024-07-23"},{"p":"nano-gpt","id":"claude-sonnet-4-5-20250929-thinking","n":"Claude Sonnet 4.5 Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"nano-gpt","id":"asi1-mini","n":"ASI1 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1,"co":1,"ctx":128000,"out":16384,"mi":"tp","mo":"t","rd":"2025-03-25","lu":"2025-03-25"},{"p":"nano-gpt","id":"qwen3.5-27b","n":"Qwen3.5 27B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.27,"co":2.16,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"Qwen3.5-27B-Omega-Evolution-v2.2-Derestricted-Lite","n":"Qwen3.5 27B Omega Evolution v2.2 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"claude-opus-4-1-thinking:32768","n":"Claude 4.1 Opus Thinking (32K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"gemma-4-31B-K1-v5","n":"Gemma 4 31B K1 v5","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"sonar","n":"Perplexity Simple","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.003,"co":1.003,"ctx":127000,"out":128000,"mi":"t","mo":"t","rd":"2025-02-19","lu":"2025-02-19"},{"p":"nano-gpt","id":"MiniMax-M2","n":"MiniMax M2","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.17,"co":1.53,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"rd":"2025-10-25","lu":"2025-10-25"},{"p":"nano-gpt","id":"sarvam-105b","n":"Sarvam 105B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.045,"co":0.177,"cr":0.028,"ctx":131072,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"Qwen3.5-27B-Writer-V2-Derestricted-Lite","n":"Qwen3.5 27B Writer V2 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"Qwen3.5-27B-Marvin-V2-Derestricted","n":"Qwen3.5 27B Marvin V2 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"claw-high","n":"Claw High","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"qwen-max","n":"Qwen 2.5 Max","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.5997,"co":6.392,"ctx":32000,"out":8192,"mi":"t","mo":"t","rd":"2024-04-03","lu":"2024-04-03"},{"p":"nano-gpt","id":"gemma-4-31B-Fabled","n":"Gemma 4 31B Fabled","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"v0-1.5-md","n":"v0 1.5 MD","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":3,"co":15,"ctx":200000,"out":64000,"mi":"t","mo":"t","rd":"2025-07-04","lu":"2025-07-04"},{"p":"nano-gpt","id":"claude-opus-4-thinking:1024","n":"Claude 4 Opus Thinking (1K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"gemini-2.5-flash-preview-04-17","n":"Gemini 2.5 Flash Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-04-17","lu":"2025-04-17"},{"p":"nano-gpt","id":"gemini-2.5-pro-exp-03-25","n":"Gemini 2.5 Pro Experimental 0325","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":10,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-03-25","lu":"2025-03-25"},{"p":"nano-gpt","id":"hermes-medium","n":"Hermes Medium","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-11","lu":"2026-05-11"},{"p":"nano-gpt","id":"sonar-pro","n":"Perplexity Pro","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":200000,"out":128000,"mi":"t","mo":"t","rd":"2025-02-19","lu":"2025-02-19"},{"p":"nano-gpt","id":"doubao-1-5-thinking-pro-250415","n":"Doubao 1.5 Thinking Pro","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.6,"co":2.4,"ctx":128000,"out":16384,"mi":"tp","mo":"t","rd":"2025-04-17","lu":"2025-04-17"},{"p":"nano-gpt","id":"qwen3.7-max:thinking","n":"Qwen3.7 Max Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":7.5,"cr":0.25,"ctx":1000000,"out":65536,"mi":"t","mo":"t","rs":1,"rd":"2026-05-21","lu":"2026-05-21"},{"p":"nano-gpt","id":"ernie-4.5-turbo-vl-32k","n":"Ernie 4.5 Turbo VL 32k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.495,"co":1.43,"ctx":32000,"out":16384,"mi":"ti","mo":"t","rd":"2025-05-08","lu":"2025-05-08"},{"p":"nano-gpt","id":"gemini-2.5-pro-preview-05-06","n":"Gemini 2.5 Pro Preview 0506","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":10,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-05-06","lu":"2025-05-06"},{"p":"nano-gpt","id":"hunyuan-turbos-20250226","n":"Hunyuan Turbo S","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.187,"co":0.374,"ctx":24000,"out":8192,"mi":"t","mo":"t","rd":"2025-02-27","lu":"2025-02-27"},{"p":"nano-gpt","id":"claude-sonnet-4-thinking:64000","n":"Claude 4 Sonnet Thinking (64K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"ernie-x1-32k","n":"Ernie X1 32k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.33,"co":1.32,"ctx":32000,"out":16384,"mi":"ti","mo":"t","rd":"2025-05-08","lu":"2025-05-08"},{"p":"nano-gpt","id":"command-a-reasoning-08-2025","n":"Cohere Command A (08/2025)","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2.5,"co":10,"ctx":256000,"out":8192,"mi":"t","mo":"t","rd":"2025-08-22","lu":"2025-08-22"},{"p":"nano-gpt","id":"doubao-seed-2-0-code-preview-260215","n":"Doubao Seed 2.0 Code Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.782,"co":3.893,"ctx":256000,"out":128000,"mi":"t","mo":"t","rd":"2026-02-14","lu":"2026-02-14"},{"p":"nano-gpt","id":"doubao-1.5-pro-256k","n":"Doubao 1.5 Pro 256k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.799,"co":1.445,"ctx":256000,"out":16384,"mi":"t","mo":"t","rd":"2025-03-12","lu":"2025-03-12"},{"p":"nano-gpt","id":"qwen3.5-35b-a3b","n":"Qwen3.5 35B A3B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.225,"co":1.8,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"qwen3.5-122b-a10b:thinking","n":"Qwen3.5 122B A10B Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.36,"co":2.88,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"yi-lightning","n":"Yi Lightning","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2006,"co":0.2006,"ctx":12000,"out":4096,"mi":"t","mo":"t","rd":"2024-10-16","lu":"2024-10-16"},{"p":"nano-gpt","id":"claude-sonnet-4-5-20250929","n":"Claude Sonnet 4.5","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.992,"co":14.994,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","tc":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"nano-gpt","id":"deepseek-math-v2","n":"DeepSeek Math V2","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.6,"co":2.2,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-12-03","lu":"2025-12-03"},{"p":"nano-gpt","id":"claude-opus-4-1-thinking:8192","n":"Claude 4.1 Opus Thinking (8K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"deepseek-reasoner","n":"DeepSeek Reasoner","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.7,"ctx":64000,"out":65536,"mi":"t","mo":"t","rd":"2025-01-20","lu":"2025-01-20"},{"p":"nano-gpt","id":"gemini-2.5-flash-nothinking","n":"Gemini 2.5 Flash (No Thinking)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":2.5,"ctx":1048756,"out":65536,"mi":"tip","mo":"t","rd":"2025-06-05","lu":"2025-06-05"},{"p":"nano-gpt","id":"Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted-Lite","n":"Qwen3.5 27B Omega Evolution v2.0 Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"fastgpt","n":"Web Answer","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":7.5,"co":7.5,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2023-08-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-Infracelestial","n":"Qwen3.5 27B Infracelestial","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"glm-4-flash","n":"GLM-4 Flash","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1003,"co":0.1003,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2024-08-01","lu":"2024-08-01"},{"p":"nano-gpt","id":"azure-gpt-4o-mini","n":"Azure gpt-4o-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1496,"co":0.595,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2024-07-18","lu":"2024-07-18"},{"p":"nano-gpt","id":"sonar-deep-research","n":"Perplexity Deep Research","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":3.4,"co":13.6,"ctx":60000,"out":128000,"mi":"t","mo":"t","rd":"2025-02-25","lu":"2025-02-25"},{"p":"nano-gpt","id":"qwq-32b","n":"Qwen: QwQ 32B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25599999,"co":0.30499999,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"mistral-code-agent-latest","n":"Mistral Code Agent Latest","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":2,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-06-02","lu":"2026-06-02"},{"p":"nano-gpt","id":"glm-4.1v-thinking-flash","n":"GLM 4.1V Thinking Flash","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.3,"co":0.3,"ctx":64000,"out":8192,"mi":"ti","mo":"t","rd":"2025-07-09","lu":"2025-07-09"},{"p":"nano-gpt","id":"qwen3.5-omni-flash","n":"Qwen3.5 Omni Flash","d":"Omni-modal model for text, vision, audio, and multimodal agent tasks","ci":0,"co":0,"ctx":49152,"out":16384,"mi":"tiva","mo":"t","rd":"2026-03-30","lu":"2026-03-30"},{"p":"nano-gpt","id":"gemma-4-31B-MeroMero","n":"Gemma 4 31B MeroMero","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"gemini-2.5-flash-preview-04-17:thinking","n":"Gemini 2.5 Flash Preview Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":3.5,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-04-17","lu":"2025-04-17"},{"p":"nano-gpt","id":"glm-4-air","n":"GLM-4 Air","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2006,"co":0.2006,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2024-06-05","lu":"2024-06-05"},{"p":"nano-gpt","id":"doubao-seed-1-6-thinking-250615","n":"Doubao Seed 1.6 Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.204,"co":2.04,"ctx":256000,"out":16384,"mi":"t","mo":"t","rd":"2025-06-15","lu":"2025-06-15"},{"p":"nano-gpt","id":"auto-model-basic","n":"Auto model (Basic)","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":9.996,"co":19.992,"ctx":1000000,"out":1000000,"mi":"t","mo":"t","rd":"2024-06-01","lu":"2024-06-01"},{"p":"nano-gpt","id":"Qwen3.5-27B-RpRMax-v1","n":"Qwen3.5 27B RpRMax v1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"ernie-5.1","n":"ERNIE 5.1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":3,"cr":0.75,"ctx":119000,"out":64000,"mi":"tiv","mo":"t","rd":"2026-05-10","lu":"2026-05-10"},{"p":"nano-gpt","id":"Qwen3.5-27B-Marvin-DPO-V2-Derestricted","n":"Qwen3.5 27B Marvin DPO V2 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"azure-gpt-4o","n":"Azure gpt-4o","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.499,"co":9.996,"ctx":128000,"out":16384,"mi":"ti","mo":"t","tc":1,"rd":"2024-05-13","lu":"2024-05-13"},{"p":"nano-gpt","id":"deepseek-chat","n":"DeepSeek V3/Deepseek Chat","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":0.25,"co":0.7,"ctx":128000,"out":8192,"mi":"tp","mo":"t","tc":1,"rd":"2025-02-27","lu":"2025-02-27"},{"p":"nano-gpt","id":"gemini-2.5-flash-preview-05-20","n":"Gemini 2.5 Flash 0520","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":1048000,"out":65536,"mi":"ti","mo":"t","rd":"2025-05-20","lu":"2025-05-20"},{"p":"nano-gpt","id":"mercury-2","n":"Mercury 2","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":0.75,"cr":0.025,"ctx":128000,"out":50000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen3.7-plus:thinking","n":"Qwen3.7 Plus Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"cr":0.04,"ctx":983616,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"nano-gpt","id":"gemini-2.0-flash-thinking-exp-1219","n":"Gemini 2.0 Flash Thinking 1219","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1003,"co":0.408,"ctx":32767,"out":8192,"mi":"t","mo":"t","rd":"2024-12-19","lu":"2024-12-19"},{"p":"nano-gpt","id":"glm-4-plus-0111","n":"GLM 4 Plus 0111","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":9.996,"co":9.996,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2025-02-19","lu":"2025-02-19"},{"p":"nano-gpt","id":"brave","n":"Brave (Answers)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":5,"co":5,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2023-03-02","lu":"2024-01-01"},{"p":"nano-gpt","id":"glm-zero-preview","n":"GLM Zero Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.802,"co":1.802,"ctx":8000,"out":4096,"mi":"t","mo":"t","rd":"2024-12-01","lu":"2024-12-01"},{"p":"nano-gpt","id":"gemini-2.5-flash-lite-preview-06-17","n":"Gemini 2.5 Flash Lite Preview","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.15,"co":0.6,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-06-17","lu":"2025-06-17"},{"p":"nano-gpt","id":"Qwen3.5-27B-Writer-Derestricted","n":"Qwen3.5 27B Writer Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"KAT-Coder-Exp-72B-1010","n":"KAT Coder Exp 72B 1010","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.2,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-10-28","lu":"2025-10-28"},{"p":"nano-gpt","id":"gemini-2.5-pro-preview-06-05","n":"Gemini 2.5 Pro Preview 0605","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":10,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-06-05","lu":"2025-06-05"},{"p":"nano-gpt","id":"Qwen3.5-27B-Musica-v1","n":"Qwen3.5 27B Musica v1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"nano-gpt","id":"MiniMax-M1","n":"MiniMax M1","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1394,"co":1.3328,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rd":"2025-06-16","lu":"2025-06-16"},{"p":"nano-gpt","id":"doubao-1-5-thinking-pro-vision-250415","n":"Doubao 1.5 Thinking Pro Vision","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.6,"co":2.4,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"qwen3.5-27b:thinking","n":"Qwen3.5 27B Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.27,"co":2.16,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"Qwen3.5-27B-Omega-Evolution-v2.0-Derestricted","n":"Qwen3.5 27B Omega Evolution v2.0 Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"Gemma-4-31B-GarnetV2","n":"Gemma 4 31B Garnet V2","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-01","lu":"2026-05-01"},{"p":"nano-gpt","id":"qwen-turbo","n":"Qwen Turbo","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.04998,"co":0.2006,"ctx":1000000,"out":8192,"mi":"t","mo":"t","rd":"2024-11-01","lu":"2024-11-01"},{"p":"nano-gpt","id":"phi-4-multimodal-instruct","n":"Phi 4 Multimodal","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.07,"co":0.11,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"mistral-small-31-24b-instruct","n":"Mistral Small 31 24b Instruct","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.3,"ctx":128000,"out":131072,"mi":"ti","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"ernie-4.5-turbo-128k","n":"Ernie 4.5 Turbo 128k","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.132,"co":0.55,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2025-05-08","lu":"2025-05-08"},{"p":"nano-gpt","id":"Qwen3.5-27B-BlueStar-Derestricted","n":"Qwen3.5 27B BlueStar Derestricted","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-06","lu":"2026-04-06"},{"p":"nano-gpt","id":"gemini-2.5-flash-lite-preview-09-2025-thinking","n":"Gemini 2.5 Flash Lite Preview (09/2025) – Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":1048756,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"doubao-seed-1-8-251215","n":"Doubao Seed 1.8","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.612,"co":6.12,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2025-12-15","lu":"2025-12-15"},{"p":"nano-gpt","id":"qwen3.6-max-preview","n":"Qwen3.6 Max Preview","f":"qwen3.6","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":1.3,"co":7.8,"ctx":245800,"out":65536,"mi":"t","mo":"t","rd":"2026-04-20","lu":"2026-04-21"},{"p":"nano-gpt","id":"exa-answer","n":"Exa (Answer)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":2.5,"co":2.5,"ctx":4096,"out":4096,"mi":"t","mo":"t","rd":"2025-06-04","lu":"2025-06-04"},{"p":"nano-gpt","id":"Baichuan4-Air","n":"Baichuan 4 Air","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.157,"co":0.157,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-08-19","lu":"2025-08-19"},{"p":"nano-gpt","id":"qwen3.5-122b-a10b","n":"Qwen3.5 122B A10B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.36,"co":2.88,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"sarvam-30b","n":"Sarvam 30B","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.028,"co":0.111,"cr":0.017,"ctx":65536,"out":4096,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"claude-opus-4-1-thinking","n":"Claude 4.1 Opus Thinking","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"claude-opus-4-thinking:32000","n":"Claude 4 Opus Thinking (32K)","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":14.994,"co":75.004,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-05-22","lu":"2025-05-22"},{"p":"nano-gpt","id":"auto-model","n":"Auto model","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0,"co":0,"ctx":1000000,"out":1000000,"mi":"t","mo":"t","rd":"2024-06-01","lu":"2024-06-01"},{"p":"nano-gpt","id":"qwen3-vl-235b-a22b-instruct-original","n":"Qwen3 VL 235B A22B Instruct Original","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.2,"ctx":32768,"out":32768,"mi":"ti","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"glm-z1-air","n":"GLM Z1 Air","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.07,"co":0.07,"ctx":32000,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"Qwen3.5-27B-Queen-Derestricted-Lite","n":"Qwen3.5 27B Queen Derestricted Lite","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.306,"co":0.306,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"inclusionai/ling-2.6-1t","n":"Ling 2.6 1T","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.3,"co":2.5,"cr":0.06,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-04-23","lu":"2026-04-23"},{"p":"nano-gpt","id":"inclusionai/ring-2.6-1t","n":"Ring 2.6 1T","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":1,"co":3,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-08","lu":"2026-05-08"},{"p":"nano-gpt","id":"inclusionai/ling-2.6-flash","n":"Ling 2.6 Flash","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.08,"co":0.24,"ctx":262144,"out":32768,"mi":"t","mo":"t","tc":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"nano-gpt","id":"Alibaba-NLP/Tongyi-DeepResearch-30B-A3B","n":"Tongyi DeepResearch 30B A3B","f":"yi","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.08,"co":0.24000000000000002,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-08-26","lu":"2025-08-26"},{"p":"nano-gpt","id":"ibm-granite/granite-4.1-8b","n":"Granite 4.1 8B","d":"Tool-capable chat model for instruction following and agentic application workflows","ci":0.05,"co":0.1,"cr":0.05,"ctx":131072,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"nano-gpt","id":"Salesforce/Llama-xLAM-2-70b-fc-r","n":"Llama-xLAM-2 70B fc-r","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.5,"co":2.5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-13","lu":"2025-04-13"},{"p":"nano-gpt","id":"THUDM/GLM-Z1-32B-0414","n":"GLM Z1 32B 0414","f":"glm-z","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.2,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"THUDM/GLM-4-32B-0414","n":"GLM 4 32B 0414","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.2,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nano-gpt","id":"THUDM/GLM-4-9B-0414","n":"GLM 4 9B 0414","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.2,"ctx":32000,"out":8000,"mi":"t","mo":"t","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nano-gpt","id":"THUDM/GLM-Z1-9B-0414","n":"GLM Z1 9B 0414","f":"glm-z","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.2,"ctx":32000,"out":8000,"mi":"t","mo":"t","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nano-gpt","id":"meta-llama/llama-3.1-8b-instruct","n":"Llama 3.1 8b Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.0544,"co":0.0544,"ctx":131072,"out":16384,"mi":"t","mo":"t","rd":"2024-07-23","lu":"2024-07-23"},{"p":"nano-gpt","id":"meta-llama/llama-4-maverick","n":"Llama 4 Maverick","f":"llama","d":"Open multimodal Llama model for strong reasoning and fast responses","ci":0.18000000000000002,"co":0.8,"ctx":1048576,"out":65536,"mi":"ti","mo":"t","tc":1,"rd":"2025-09-05","lu":"2025-09-05"},{"p":"nano-gpt","id":"meta-llama/llama-3.3-70b-instruct","n":"Llama 3.3 70b Instruct","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.05,"co":0.23,"ctx":131072,"out":16384,"mi":"t","mo":"t","tc":1,"rd":"2025-02-27","lu":"2025-02-27"},{"p":"nano-gpt","id":"meta-llama/llama-4-scout","n":"Llama 4 Scout","f":"llama","d":"Open multimodal Llama model for long-context analysis and efficient agents","ci":0.085,"co":0.46,"ctx":328000,"out":65536,"mi":"ti","mo":"t","tc":1,"rd":"2025-09-05","lu":"2025-09-05"},{"p":"nano-gpt","id":"meta-llama/llama-3.2-3b-instruct","n":"Llama 3.2 3b Instruct","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.0306,"co":0.0493,"ctx":131072,"out":8192,"mi":"tp","mo":"t","rd":"2024-09-25","lu":"2024-09-25"},{"p":"nano-gpt","id":"featherless-ai/Qwerky-72B","n":"Qwerky 72B","f":"qwerky","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.5,"co":0.5,"ctx":32000,"out":8192,"mi":"t","mo":"t","rd":"2025-03-20","lu":"2025-03-20"},{"p":"nano-gpt","id":"moonshotai/kimi-k2-instruct-0711","n":"Kimi K2 0711","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.1,"co":2,"ctx":128000,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-07-11","lu":"2025-07-11"},{"p":"nano-gpt","id":"moonshotai/Kimi-K2-Instruct-0905","n":"Kimi K2 0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.4,"co":2,"ctx":256000,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"moonshotai/kimi-k2-thinking-original","n":"Kimi K2 Thinking Original","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.6,"co":2.5,"ctx":256000,"out":16384,"mi":"t","mo":"t","rs":1,"rd":"2025-11-06","lu":"2025-11-06"},{"p":"nano-gpt","id":"moonshotai/kimi-k2.5:thinking","n":"Kimi K2.5 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.3,"co":1.9,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-01-26","lu":"2026-01-26"},{"p":"nano-gpt","id":"moonshotai/kimi-k2-instruct","n":"Kimi K2 Instruct","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.1,"co":2,"ctx":256000,"out":8192,"mi":"t","mo":"t","tc":1,"rd":"2025-07-01","lu":"2025-07-01"},{"p":"nano-gpt","id":"moonshotai/kimi-k2-thinking","n":"Kimi K2 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.3,"co":1.2,"ctx":256000,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2025-11-06","lu":"2025-11-06"},{"p":"nano-gpt","id":"moonshotai/kimi-k2.6:thinking","n":"Kimi K2.6 Thinking","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.53,"co":2.73,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-16","lu":"2026-04-21"},{"p":"nano-gpt","id":"moonshotai/kimi-k2.5","n":"Kimi K2.5","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.3,"co":1.9,"ctx":256000,"out":65536,"mi":"ti","mo":"t","tc":1,"rd":"2026-01-26","lu":"2026-01-26"},{"p":"nano-gpt","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.53,"co":2.73,"ctx":256000,"out":65536,"mi":"ti","mo":"t","tc":1,"ow":1,"rd":"2026-04-16","lu":"2026-04-21"},{"p":"nano-gpt","id":"moonshotai/kimi-latest","n":"Kimi Latest","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.5,"co":2.6,"cr":0.125,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-03","lu":"2026-05-03"},{"p":"nano-gpt","id":"moonshotai/kimi-k2-thinking-turbo-original","n":"Kimi K2 Thinking Turbo Original","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":1.15,"co":8,"ctx":256000,"out":16384,"mi":"t","mo":"t","rs":1,"rd":"2025-11-06","lu":"2025-11-06"},{"p":"nano-gpt","id":"baidu/ernie-4.5-vl-28b-a3b","n":"ERNIE 4.5 VL 28B","f":"ernie","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.13999999999999999,"co":0.5599999999999999,"ctx":32768,"out":16384,"mi":"ti","mo":"t","rd":"2025-06-30","lu":"2025-06-30"},{"p":"nano-gpt","id":"perceptron/perceptron-mk1","n":"Perceptron Mk1","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.15,"co":1.5,"ctx":32768,"out":8192,"mi":"tiv","mo":"t","rs":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"failspy/Meta-Llama-3-70B-Instruct-abliterated-v3.5","n":"Llama 3 70B abliterated","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.7,"co":0.7,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"nanogpt/coding-router:max","n":"Coding Router Max","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"nanogpt/coding-router:high","n":"Coding Router High","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":1.1,"co":2.2,"cr":0.11,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"nanogpt/coding-router:low","n":"Coding Router Low","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"nanogpt/coding-router:medium","n":"Coding Router Medium","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":0.14,"co":0.28,"cr":0.028,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"nanogpt/coding-router","n":"Coding Router","d":"Automatic model router for matching prompts to suitable backends and budgets","ci":1.1,"co":2.2,"cr":0.11,"ctx":1000000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-12","lu":"2026-05-12"},{"p":"nano-gpt","id":"GalrionSoftworks/MN-LooseCannon-12B-v1","n":"MN-LooseCannon-12B-v1","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"TheDrummer/Cydonia-24B-v4.3","n":"The Drummer Cydonia 24B v4.3","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1003,"co":0.1207,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-12-25","lu":"2025-12-25"},{"p":"nano-gpt","id":"TheDrummer/Cydonia-24B-v4.1","n":"The Drummer Cydonia 24B v4.1","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1003,"co":0.1207,"ctx":16384,"out":32768,"mi":"t","mo":"t","rd":"2025-08-19","lu":"2025-08-19"},{"p":"nano-gpt","id":"TheDrummer/Cydonia-24B-v2","n":"The Drummer Cydonia 24B v2","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1003,"co":0.1207,"ctx":16384,"out":32768,"mi":"t","mo":"t","rd":"2025-02-17","lu":"2025-02-17"},{"p":"nano-gpt","id":"TheDrummer/Anubis-70B-v1.1","n":"Anubis 70B v1.1","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.31,"co":0.31,"ctx":131072,"out":16384,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"TheDrummer/UnslopNemo-12B-v4.1","n":"UnslopNemo 12b v4","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.493,"co":0.493,"ctx":32768,"out":8192,"mi":"tp","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"TheDrummer/skyfall-36b-v2","n":"TheDrummer Skyfall 36B V2","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.493,"co":0.493,"ctx":64000,"out":32768,"mi":"tp","mo":"t","rd":"2025-03-10","lu":"2025-03-10"},{"p":"nano-gpt","id":"TheDrummer/Rocinante-12B-v1.1","n":"Rocinante 12b","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.408,"co":0.595,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"TheDrummer/Magidonia-24B-v4.3","n":"The Drummer Magidonia 24B v4.3","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1003,"co":0.1207,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-12-25","lu":"2025-12-25"},{"p":"nano-gpt","id":"TheDrummer/Cydonia-24B-v4","n":"The Drummer Cydonia 24B v4","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.2006,"co":0.2414,"ctx":16384,"out":32768,"mi":"t","mo":"t","rd":"2025-07-22","lu":"2025-07-22"},{"p":"nano-gpt","id":"TheDrummer/Skyfall-31B-v4.2","n":"TheDrummer Skyfall 31B v4.2","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.55,"co":0.8,"ctx":131072,"out":16384,"mi":"tp","mo":"t","rd":"2026-03-26","lu":"2026-03-26"},{"p":"nano-gpt","id":"TheDrummer/Anubis-70B-v1","n":"Anubis 70B v1","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.31,"co":0.31,"ctx":65536,"out":16384,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"huihui-ai/Qwen2.5-32B-Instruct-abliterated","n":"Qwen 2.5 32B Abliterated","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.7,"co":0.7,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2025-01-06","lu":"2025-01-06"},{"p":"nano-gpt","id":"huihui-ai/Llama-3.3-70B-Instruct-abliterated","n":"Llama 3.3 70B Instruct abliterated","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.7,"co":0.7,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2025-08-08","lu":"2025-08-08"},{"p":"nano-gpt","id":"huihui-ai/DeepSeek-R1-Distill-Qwen-32B-abliterated","n":"DeepSeek R1 Qwen Abliterated","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":1.4,"co":1.4,"ctx":16384,"out":8192,"mi":"t","mo":"t","rs":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"nano-gpt","id":"huihui-ai/DeepSeek-R1-Distill-Llama-70B-abliterated","n":"DeepSeek R1 Llama 70B Abliterated","f":"deepseek","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.7,"co":0.7,"ctx":16384,"out":8192,"mi":"t","mo":"t","rs":1,"rd":"2025-01-20","lu":"2025-01-20"},{"p":"nano-gpt","id":"Envoid/Llama-3.05-NT-Storybreaker-Ministral-70B","n":"Llama 3.05 Storybreaker Ministral 70b","f":"llama","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-12-01","lu":"2024-12-01"},{"p":"nano-gpt","id":"Envoid/Llama-3.05-Nemotron-Tenyxchat-Storybreaker-70B","n":"Nemotron Tenyxchat Storybreaker 70b","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-12-01","lu":"2024-12-01"},{"p":"nano-gpt","id":"stepfun-ai/step-3.5-flash-2603","n":"Step 3.5 Flash 2603","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.1,"co":0.3,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"rd":"2026-04-14","lu":"2026-04-14"},{"p":"nano-gpt","id":"stepfun-ai/step-3.5-flash","n":"Step 3.5 Flash","f":"step","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.2,"co":0.5,"ctx":256000,"out":256000,"mi":"t","mo":"t","rs":1,"rd":"2026-02-02","lu":"2026-02-02"},{"p":"nano-gpt","id":"mistral/mistral-medium-3.5","n":"Mistral Medium 3.5","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":1.5,"co":7.5,"ctx":256000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"nano-gpt","id":"mistral/mistral-medium-3.5:thinking","n":"Mistral Medium 3.5 Thinking","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":1.5,"co":7.5,"ctx":256000,"out":32768,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"Tongyi-Zhiwen/QwenLong-L1-32B","n":"QwenLong L1 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.13999999999999999,"co":0.6,"ctx":128000,"out":40960,"mi":"t","mo":"t","rd":"2025-01-25","lu":"2025-01-25"},{"p":"nano-gpt","id":"google/gemini-flash-1.5","n":"Gemini 1.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.0748,"co":0.306,"ctx":2000000,"out":8192,"mi":"t","mo":"t","rd":"2024-05-14","lu":"2024-05-14"},{"p":"nano-gpt","id":"google/gemini-3.1-pro-preview-high","n":"Gemini 3.1 Pro (Preview High)","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-21","lu":"2026-02-21"},{"p":"nano-gpt","id":"google/gemini-3.1-flash-lite","n":"Gemini 3.1 Flash Lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-03","lu":"2026-03-03"},{"p":"nano-gpt","id":"google/gemini-3-flash-preview-thinking","n":"Gemini 3 Flash Thinking","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2025-12-17","lu":"2025-12-17"},{"p":"nano-gpt","id":"google/gemini-3.5-flash","n":"Gemini 3.5 Flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"tia","mo":"t","rs":1,"tc":1,"rd":"2026-05-19","lu":"2026-05-19"},{"p":"nano-gpt","id":"google/gemma-4-31b-it","n":"Gemma 4 31B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.1,"co":0.35,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"nano-gpt","id":"google/gemini-pro-latest","n":"Gemini Pro Latest","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-29","lu":"2026-03-29"},{"p":"nano-gpt","id":"google/gemini-3.1-pro-preview-customtools","n":"Gemini 3.1 Pro (Preview Custom Tools)","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-27","lu":"2026-02-27"},{"p":"nano-gpt","id":"google/gemini-flash-lite-latest","n":"Gemini Flash Lite Latest","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-29","lu":"2026-03-29"},{"p":"nano-gpt","id":"google/gemma-4-26b-a4b-it:thinking","n":"Gemma 4 26B A4B Thinking","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.13,"co":0.4,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"nano-gpt","id":"google/gemini-3.1-pro-preview","n":"Gemini 3.1 Pro (Preview)","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-19","lu":"2026-02-19"},{"p":"nano-gpt","id":"google/gemma-4-26b-a4b-it","n":"Gemma 4 26B A4B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.13,"co":0.4,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"nano-gpt","id":"google/gemini-3.1-pro-preview-low","n":"Gemini 3.1 Pro (Preview Low)","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-02-21","lu":"2026-02-21"},{"p":"nano-gpt","id":"google/gemma-4-31b-it:thinking","n":"Gemma 4 31B Thinking","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.1,"co":0.35,"ctx":262144,"out":131072,"mi":"ti","mo":"t","rs":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"nano-gpt","id":"google/gemini-3.5-flash-thinking","n":"Gemini 3.5 Flash Thinking","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048576,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2026-05-19","lu":"2026-05-19"},{"p":"nano-gpt","id":"google/gemini-3-flash-preview","n":"Gemini 3 Flash (Preview)","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":0.5,"co":3,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2025-12-17","lu":"2025-12-17"},{"p":"nano-gpt","id":"google/gemini-flash-latest","n":"Gemini Flash Latest","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":1048756,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2026-03-29","lu":"2026-03-29"},{"p":"nano-gpt","id":"liquid/lfm-2-24b-a2b","n":"LFM2 24B A2B","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.03,"co":0.12,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-12-20","lu":"2025-12-20"},{"p":"nano-gpt","id":"x-ai/grok-4.20","n":"Grok 4.20","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":2000000,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-31","lu":"2026-03-31"},{"p":"nano-gpt","id":"x-ai/grok-4.3","n":"Grok 4.3","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-30","lu":"2026-04-30"},{"p":"nano-gpt","id":"x-ai/grok-4.20-multi-agent","n":"Grok 4.20 Multi-Agent","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":2,"co":6,"ctx":2000000,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-31","lu":"2026-03-31"},{"p":"nano-gpt","id":"x-ai/grok-latest","n":"Grok Latest","d":"Grok model for agentic tool use, reasoning, coding, and live assistance","ci":1.25,"co":2.5,"cr":0.2,"ctx":1000000,"out":1000000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-03","lu":"2026-05-03"},{"p":"nano-gpt","id":"x-ai/grok-build-0.1","n":"Grok Build 0.1","d":"Grok coding model for agentic engineering, edits, and codebase workflows","ci":1,"co":2,"cr":0.2,"ctx":256000,"out":256000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-20","lu":"2026-05-20"},{"p":"nano-gpt","id":"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0","n":"EVA Llama 3.33 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.006,"co":2.006,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"EVA-UNIT-01/EVA-Qwen2.5-72B-v0.2","n":"EVA-Qwen2.5-72B-v0.2","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.7989999999999999,"co":0.7989999999999999,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1","n":"EVA-LLaMA-3.33-70B-v0.1","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.006,"co":2.006,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"EVA-UNIT-01/EVA-Qwen2.5-32B-v0.2","n":"EVA-Qwen2.5-32B-v0.2","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.7989999999999999,"co":0.7989999999999999,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"microsoft/wizardlm-2-8x22b","n":"WizardLM-2 8x22B","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":65536,"out":8192,"mi":"tp","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"Doctor-Shotgun/MS3.2-24B-Magnum-Diamond","n":"MS3.2 24B Magnum Diamond","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":32768,"mi":"t","mo":"t","rd":"2025-11-24","lu":"2025-11-24"},{"p":"nano-gpt","id":"poolside/laguna-xs.2","n":"Laguna XS.2","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1,"co":0.3,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2026-04-29","lu":"2026-04-29"},{"p":"nano-gpt","id":"poolside/laguna-m.1","n":"Laguna M.1","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.1,"co":0.3,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2026-04-29","lu":"2026-04-29"},{"p":"nano-gpt","id":"z-ai/glm-4.5v:thinking","n":"GLM 4.5V Thinking","f":"glmv","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.7999999999999998,"ctx":64000,"out":96000,"mi":"ti","mo":"t","rs":1,"rd":"2025-11-22","lu":"2025-11-22"},{"p":"nano-gpt","id":"z-ai/glm-4.6:thinking","n":"GLM 4.6 Thinking","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.4,"co":1.5,"ctx":200000,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"nano-gpt","id":"z-ai/glm-4.5v","n":"GLM 4.5V","f":"glmv","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":1.7999999999999998,"ctx":64000,"out":96000,"mi":"ti","mo":"t","rs":1,"rd":"2025-11-22","lu":"2025-11-22"},{"p":"nano-gpt","id":"z-ai/glm-4.6","n":"GLM 4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.4,"co":1.5,"ctx":200000,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-09-30","lu":"2025-09-30"},{"p":"nano-gpt","id":"z-ai/glm-5v-turbo:thinking","n":"GLM 5V Turbo Thinking","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":1.2,"co":4,"cr":0.24,"ctx":202800,"out":131100,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"nano-gpt","id":"z-ai/glm-5v-turbo","n":"GLM 5V Turbo","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":1.2,"co":4,"cr":0.24,"ctx":202800,"out":131100,"mi":"ti","mo":"t","tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"nano-gpt","id":"z-ai/glm-5-turbo","n":"GLM 5 Turbo","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":1.2,"co":4,"cr":0.24,"ctx":202800,"out":131072,"mi":"t","mo":"t","tc":1,"rd":"2026-03-15","lu":"2026-03-15"},{"p":"nano-gpt","id":"openai/o3-mini-low","n":"OpenAI o3-mini (Low)","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":9.996,"co":19.992,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-31","lu":"2025-01-31"},{"p":"nano-gpt","id":"openai/gpt-oss-safeguard-20b","n":"GPT OSS Safeguard 20B","f":"gpt-oss","d":"Safety model for policy screening, moderation, and risk-aware routing workflows","ci":0.075,"co":0.3,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"rd":"2025-10-29","lu":"2025-10-29"},{"p":"nano-gpt","id":"openai/o3","n":"OpenAI o3","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":2,"co":8,"ctx":200000,"out":100000,"mi":"t","mo":"t","rd":"2025-04-16","lu":"2025-04-16"},{"p":"nano-gpt","id":"openai/o4-mini-high","n":"OpenAI o4-mini high","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"nano-gpt","id":"openai/o3-pro-2025-06-10","n":"OpenAI o3-pro (2025-06-10)","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":9.996,"co":19.992,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-06-10","lu":"2025-06-10"},{"p":"nano-gpt","id":"openai/gpt-5.2-pro","n":"GPT 5.2 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":21,"co":168,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-01-01","lu":"2026-01-01"},{"p":"nano-gpt","id":"openai/gpt-4o-mini-search-preview","n":"GPT-4o mini Search Preview","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.088,"co":0.35,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2024-07-18","lu":"2024-07-18"},{"p":"nano-gpt","id":"openai/gpt-5","n":"GPT 5","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-08-07","lu":"2025-08-07"},{"p":"nano-gpt","id":"openai/gpt-3.5-turbo","n":"GPT-3.5 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.5,"co":1.5,"ctx":16385,"out":4096,"mi":"t","mo":"t","rd":"2022-11-30","lu":"2024-01-01"},{"p":"nano-gpt","id":"openai/gpt-5-pro","n":"GPT 5 Pro","f":"gpt-pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":15,"co":120,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2025-08-07","lu":"2025-08-07"},{"p":"nano-gpt","id":"openai/gpt-4o","n":"GPT-4o","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.499,"co":9.996,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-05-13","lu":"2024-05-13"},{"p":"nano-gpt","id":"openai/o4-mini","n":"OpenAI o4-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"nano-gpt","id":"openai/gpt-5.4-nano","n":"GPT 5.4 Nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-17","lu":"2026-03-17"},{"p":"nano-gpt","id":"openai/gpt-5.1-codex","n":"GPT 5.1 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"nano-gpt","id":"openai/gpt-5.1-codex-max","n":"GPT 5.1 Codex Max","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":2.5,"co":20,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"nano-gpt","id":"openai/gpt-4o-2024-08-06","n":"GPT-4o (2024-08-06)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.499,"co":9.996,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-08-06","lu":"2024-08-06"},{"p":"nano-gpt","id":"openai/o1-preview","n":"OpenAI o1-preview","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":14.993999999999998,"co":59.993,"ctx":128000,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2024-09-12","lu":"2024-09-12"},{"p":"nano-gpt","id":"openai/o3-mini","n":"OpenAI o3-mini","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":1.1,"co":4.4,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-31","lu":"2025-01-31"},{"p":"nano-gpt","id":"openai/gpt-5.2","n":"GPT 5.2","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-01-01","lu":"2026-01-01"},{"p":"nano-gpt","id":"openai/gpt-5.3-codex","n":"GPT 5.3 Codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-24","lu":"2026-02-24"},{"p":"nano-gpt","id":"openai/gpt-latest","n":"GPT Latest","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-29","lu":"2026-03-29"},{"p":"nano-gpt","id":"openai/gpt-5.1-codex-mini","n":"GPT 5.1 Codex Mini","f":"gpt-codex-mini","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":0.25,"co":2,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"nano-gpt","id":"openai/o4-mini-deep-research","n":"OpenAI o4-mini Deep Research","f":"o-mini","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":9.996,"co":19.992,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"nano-gpt","id":"openai/gpt-4.1-nano","n":"GPT 4.1 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1,"co":0.4,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nano-gpt","id":"openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.25,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"nano-gpt","id":"openai/gpt-4o-2024-11-20","n":"GPT-4o (2024-11-20)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.5,"co":10,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-11-20","lu":"2024-11-20"},{"p":"nano-gpt","id":"openai/o1","n":"OpenAI o1","f":"o","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":14.993999999999998,"co":59.993,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"rd":"2024-12-17","lu":"2024-12-17"},{"p":"nano-gpt","id":"openai/o1-pro","n":"OpenAI o1 Pro","f":"o-pro","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":150,"co":600,"ctx":200000,"out":100000,"mi":"tip","mo":"t","rd":"2025-01-25","lu":"2025-01-25"},{"p":"nano-gpt","id":"openai/gpt-chat-latest","n":"GPT Chat Latest","d":"Chat-tuned GPT model for conversational assistance, writing, and tool workflows","ci":5,"co":30,"cr":0.5,"ctx":400000,"out":128000,"mi":"tip","mo":"t","tc":1,"rd":"2026-05-03","lu":"2026-05-03"},{"p":"nano-gpt","id":"openai/gpt-5.4","n":"GPT 5.4","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":2.5,"co":15,"cr":0.25,"ctx":922000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-05","lu":"2026-03-05"},{"p":"nano-gpt","id":"openai/gpt-5.4-mini","n":"GPT 5.4 Mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-17","lu":"2026-03-17"},{"p":"nano-gpt","id":"openai/gpt-4.1","n":"GPT 4.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2,"co":8,"ctx":1047576,"out":32768,"mi":"tip","mo":"t","tc":1,"rd":"2025-09-10","lu":"2025-09-10"},{"p":"nano-gpt","id":"openai/o3-deep-research","n":"OpenAI o3 Deep Research","f":"o","d":"Research model for long-horizon investigation, synthesis, and analytical reports","ci":9.996,"co":19.992,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"rd":"2025-04-16","lu":"2025-04-16"},{"p":"nano-gpt","id":"openai/gpt-4-turbo-preview","n":"GPT-4 Turbo Preview","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":9.996,"co":30.004999999999995,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2023-11-06","lu":"2024-01-01"},{"p":"nano-gpt","id":"openai/gpt-5-mini","n":"GPT 5 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.25,"co":2,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2025-08-07","lu":"2025-08-07"},{"p":"nano-gpt","id":"openai/gpt-4.1-mini","n":"GPT 4.1 Mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.4,"co":1.6,"ctx":1047576,"out":32768,"mi":"ti","mo":"t","rd":"2025-04-14","lu":"2025-04-14"},{"p":"nano-gpt","id":"openai/gpt-4-turbo","n":"GPT-4 Turbo","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":10,"co":30,"ctx":128000,"out":4096,"mi":"ti","mo":"t","rd":"2023-11-06","lu":"2024-01-01"},{"p":"nano-gpt","id":"openai/gpt-5-nano","n":"GPT 5 Nano","f":"gpt-nano","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.05,"co":0.4,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"rd":"2025-08-07","lu":"2025-08-07"},{"p":"nano-gpt","id":"openai/gpt-5.4-pro","n":"GPT 5.4 Pro","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":30,"co":180,"cr":3,"ctx":922000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-05","lu":"2026-03-05"},{"p":"nano-gpt","id":"openai/o3-mini-high","n":"OpenAI o3-mini (High)","f":"o-mini","d":"O-series reasoning model for hard analysis, math, coding, and planning","ci":0.64,"co":2.588,"ctx":200000,"out":100000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-01-31","lu":"2025-01-31"},{"p":"nano-gpt","id":"openai/gpt-4o-search-preview","n":"GPT-4o Search Preview","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.47,"co":5.88,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-05-13","lu":"2024-05-13"},{"p":"nano-gpt","id":"openai/gpt-5.1-2025-11-13","n":"GPT-5.1 (2025-11-13)","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":1000000,"out":32768,"mi":"t","mo":"t","rd":"2025-11-13","lu":"2025-11-13"},{"p":"nano-gpt","id":"openai/gpt-4o-mini","n":"GPT-4o mini","f":"gpt-mini","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.1496,"co":0.595,"ctx":128000,"out":16384,"mi":"ti","mo":"t","rd":"2024-07-18","lu":"2024-07-18"},{"p":"nano-gpt","id":"openai/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.04,"co":0.15,"ctx":128000,"out":16384,"mi":"t","mo":"t","rs":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"nano-gpt","id":"openai/gpt-5-codex","n":"GPT-5 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":9.996,"co":19.992,"ctx":256000,"out":32768,"mi":"t","mo":"t","rd":"2025-09-15","lu":"2025-09-15"},{"p":"nano-gpt","id":"openai/gpt-5.2-codex","n":"GPT 5.2 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-01-14","lu":"2026-01-14"},{"p":"nano-gpt","id":"openai/gpt-5.1","n":"GPT 5.1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":1.25,"co":10,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2025-11-13","lu":"2025-11-13"},{"p":"nano-gpt","id":"openai/gpt-5.5","n":"GPT 5.5","d":"Frontier GPT model for professional reasoning, coding, and multimodal work","ci":5,"co":30,"cr":0.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-23","lu":"2026-04-23"},{"p":"nano-gpt","id":"Gryphe/MythoMax-L2-13b","n":"MythoMax 13B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.1003,"co":0.1003,"ctx":4000,"out":4096,"mi":"t","mo":"t","rd":"2025-08-08","lu":"2025-08-08"},{"p":"nano-gpt","id":"Unbabel/M-Prometheus-14B","n":"M-Prometheus 14B","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.2,"co":0.2,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2026-05-29","lu":"2026-05-29"},{"p":"nano-gpt","id":"LLM360/K2-Think","n":"K2-Think","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.17,"co":0.68,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"NousResearch/hermes-4-405b","n":"Hermes 4 Large","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.3,"co":1.2,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2025-08-26","lu":"2025-08-26"},{"p":"nano-gpt","id":"NousResearch/hermes-3-llama-3.1-70b","n":"Hermes 3 70B","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.408,"co":0.408,"ctx":65536,"out":8192,"mi":"t","mo":"t","rd":"2026-01-07","lu":"2026-01-07"},{"p":"nano-gpt","id":"NousResearch/Hermes-4-70B:thinking","n":"Hermes 4 (Thinking)","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.2006,"co":0.3995,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2025-09-17","lu":"2025-09-17"},{"p":"nano-gpt","id":"NousResearch/hermes-4-70b","n":"Hermes 4 Medium","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.2006,"co":0.3995,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2025-07-03","lu":"2025-07-03"},{"p":"nano-gpt","id":"NousResearch/DeepHermes-3-Mistral-24B-Preview","n":"DeepHermes-3 Mistral 24B (Preview)","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.3,"co":0.3,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-05-10","lu":"2025-05-10"},{"p":"nano-gpt","id":"NousResearch/hermes-4-405b:thinking","n":"Hermes 4 Large (Thinking)","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.3,"co":1.2,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"unsloth/gemma-3-12b-it","n":"Gemma 3 12B IT","f":"unsloth","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.272,"co":0.272,"ctx":128000,"out":131072,"mi":"tp","mo":"t","rd":"2025-03-10","lu":"2025-03-10"},{"p":"nano-gpt","id":"unsloth/gemma-3-4b-it","n":"Gemma 3 4B IT","f":"unsloth","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.2006,"co":0.2006,"ctx":128000,"out":8192,"mi":"tp","mo":"t","rd":"2025-03-10","lu":"2025-03-10"},{"p":"nano-gpt","id":"unsloth/gemma-3-27b-it","n":"Gemma 3 27B IT","f":"unsloth","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.2992,"co":0.2992,"ctx":128000,"out":96000,"mi":"tp","mo":"t","rd":"2025-03-10","lu":"2025-03-10"},{"p":"nano-gpt","id":"NeverSleep/Lumimaid-v0.2-70B","n":"Lumimaid v0.2","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":1,"co":1.5,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"mistralai/mixtral-8x7b-instruct-v0.1","n":"Mixtral 8x7B","f":"mixtral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.27,"co":0.27,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-12-11","lu":"2025-12-11"},{"p":"nano-gpt","id":"mistralai/mistral-small-4-119b-2603:thinking","n":"Mistral Small 4 119B Thinking","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.4,"co":1.4,"ctx":262144,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-17","lu":"2026-03-17"},{"p":"nano-gpt","id":"mistralai/mixtral-8x22b-instruct-v0.1","n":"Mixtral 8x22B","f":"mixtral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.8999999999999999,"co":0.8999999999999999,"ctx":65536,"out":32768,"mi":"t","mo":"t","rd":"2025-12-11","lu":"2025-12-11"},{"p":"nano-gpt","id":"mistralai/Devstral-Small-2505","n":"Mistral Devstral Small 2505","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.060000000000000005,"co":0.060000000000000005,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2025-08-02","lu":"2025-08-02"},{"p":"nano-gpt","id":"mistralai/ministral-8b-2512","n":"Ministral 8B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.15,"co":0.15,"ctx":262144,"out":32768,"mi":"t","mo":"t","rd":"2025-12-04","lu":"2025-12-04"},{"p":"nano-gpt","id":"mistralai/mistral-saba","n":"Mistral Saba","f":"mistral","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.1989,"co":0.595,"ctx":32000,"out":32768,"mi":"t","mo":"t","rd":"2025-02-17","lu":"2025-02-17"},{"p":"nano-gpt","id":"mistralai/mistral-large","n":"Mistral Large 2411","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":2.006,"co":6.001,"ctx":128000,"out":256000,"mi":"t","mo":"t","rd":"2024-02-26","lu":"2024-02-26"},{"p":"nano-gpt","id":"mistralai/mistral-medium-3.1","n":"Mistral Medium 3.1","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rd":"2025-09-05","lu":"2025-09-05"},{"p":"nano-gpt","id":"mistralai/ministral-3b-2512","n":"Ministral 3B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.1,"ctx":131072,"out":32768,"mi":"t","mo":"t","rd":"2025-12-04","lu":"2025-12-04"},{"p":"nano-gpt","id":"mistralai/ministral-14b-instruct-2512","n":"Ministral 3 14B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.1,"co":0.4,"ctx":262144,"out":32768,"mi":"ti","mo":"t","rd":"2025-12-02","lu":"2025-12-02"},{"p":"nano-gpt","id":"mistralai/mistral-small-4-119b-2603","n":"Mistral Small 4 119B","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.4,"co":1.4,"ctx":262144,"out":16384,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"nano-gpt","id":"mistralai/ministral-14b-2512","n":"Ministral 14B","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":0.2,"co":0.2,"ctx":262144,"out":32768,"mi":"t","mo":"t","rd":"2025-12-04","lu":"2025-12-04"},{"p":"nano-gpt","id":"mistralai/mistral-large-3-675b-instruct-2512","n":"Mistral Large 3 675B","f":"mistral-large","d":"Flagship Mistral model for advanced reasoning, coding, and multilingual work","ci":1,"co":3,"ctx":262144,"out":256000,"mi":"ti","mo":"t","rd":"2025-12-02","lu":"2025-12-02"},{"p":"nano-gpt","id":"mistralai/mistral-medium-3","n":"Mistral Medium 3","f":"mistral-medium","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.4,"co":2,"ctx":131072,"out":32768,"mi":"ti","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"mistralai/devstral-2-123b-instruct-2512","n":"Devstral 2 123B","f":"devstral","d":"Mistral coding agent model for repository tasks and software engineering workflows","ci":0.4,"co":1.4,"ctx":262144,"out":65536,"mi":"t","mo":"t","rd":"2025-12-09","lu":"2025-12-09"},{"p":"nano-gpt","id":"mistralai/codestral-2508","n":"Codestral 2508","f":"codestral","d":"Mistral coding model for code completion, generation, and developer workflows","ci":0.3,"co":0.8999999999999999,"ctx":256000,"out":32768,"mi":"t","mo":"t","rd":"2025-08-01","lu":"2025-08-01"},{"p":"nano-gpt","id":"mistralai/Mistral-Nemo-Instruct-2407","n":"Mistral Nemo","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.1003,"co":0.1207,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-18","lu":"2024-07-18"},{"p":"nano-gpt","id":"bytedance-seed/seed-2.0-lite","n":"ByteDance Seed 2.0 Lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.25,"co":2,"ctx":262144,"out":131072,"mi":"t","mo":"t","rd":"2026-03-10","lu":"2026-03-10"},{"p":"nano-gpt","id":"anthracite-org/magnum-v2-72b","n":"Magnum V2 72B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2.006,"co":2.992,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"anthracite-org/magnum-v4-72b","n":"Magnum v4 72B","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":2.006,"co":2.992,"ctx":16384,"out":8192,"mi":"tp","mo":"t","rd":"2025-01-01","lu":"2025-01-01"},{"p":"nano-gpt","id":"inflatebot/MN-12B-Mag-Mell-R1","n":"Mag Mell R1","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning","n":"Nvidia Nemotron 3 Nano Omni","f":"nemotron","d":"Open Nemotron omni model combining reasoning with text, vision, and audio","ci":0.105,"co":0.42,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-04-28","lu":"2026-04-28"},{"p":"nano-gpt","id":"nvidia/Llama-3.1-Nemotron-70B-Instruct-HF","n":"Nvidia Nemotron 70b","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.357,"co":0.408,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"nvidia/nvidia-nemotron-nano-9b-v2","n":"Nvidia Nemotron Nano 9B v2","f":"nemotron","d":"Compact Nemotron model for efficient reasoning and deployable AI agents","ci":0.17,"co":0.68,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-08-18","lu":"2025-08-18"},{"p":"nano-gpt","id":"nvidia/Llama-3_3-Nemotron-Super-49B-v1_5","n":"Nvidia Nemotron Super 49B v1.5","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.05,"co":0.25,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-08-08","lu":"2025-08-08"},{"p":"nano-gpt","id":"nvidia/nemotron-3-nano-30b-a3b","n":"Nvidia Nemotron 3 Nano 30B","f":"nemotron","d":"Small Nemotron 3 MoE for efficient coding, math, and long-context agents","ci":0.17,"co":0.68,"ctx":256000,"out":262144,"mi":"t","mo":"t","rd":"2025-12-15","lu":"2025-12-15"},{"p":"nano-gpt","id":"nvidia/Llama-3.3-Nemotron-Super-49B-v1","n":"Nvidia Nemotron Super 49B","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.15,"co":0.15,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-08-08","lu":"2025-08-08"},{"p":"nano-gpt","id":"nvidia/nemotron-3-super-120b-a12b:thinking","n":"Nvidia Nemotron 3 Super 120B Thinking","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.05,"co":0.25,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"nano-gpt","id":"nvidia/nemotron-3-super-120b-a12b","n":"Nvidia Nemotron 3 Super 120B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.05,"co":0.25,"ctx":262144,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"nano-gpt","id":"cognitivecomputations/dolphin-2.9.2-qwen2-72b","n":"Dolphin 72b","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.306,"co":0.306,"ctx":8192,"out":4096,"mi":"t","mo":"t","rd":"2025-02-27","lu":"2025-02-27"},{"p":"nano-gpt","id":"xiaomi/mimo-v2-flash-original","n":"MiMo V2 Flash Original","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.102,"co":0.306,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-12-17","lu":"2025-12-17"},{"p":"nano-gpt","id":"xiaomi/mimo-v2-flash-thinking-original","n":"MiMo V2 Flash (Thinking) Original","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.102,"co":0.306,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-12-17","lu":"2025-12-17"},{"p":"nano-gpt","id":"xiaomi/mimo-v2.5","n":"MiMo V2.5","d":"MiMo omni model for text, image, video, audio, and agents","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1048576,"out":131072,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"nano-gpt","id":"xiaomi/mimo-v2-omni","n":"MiMo V2 Omni","d":"MiMo omni model for text, image, video, audio, and agents","ci":0.4,"co":2,"cr":0.08,"ctx":262144,"out":65536,"mi":"tiva","mo":"t","rs":1,"tc":1,"rd":"2026-03-19","lu":"2026-03-19"},{"p":"nano-gpt","id":"xiaomi/mimo-v2-flash","n":"MiMo V2 Flash","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.102,"co":0.306,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-12-17","lu":"2025-12-17"},{"p":"nano-gpt","id":"xiaomi/mimo-v2-pro","n":"MiMo V2 Pro","d":"MiMo pro model for strong multimodal reasoning and agent execution","ci":1,"co":3,"cr":0.2,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-19","lu":"2026-03-19"},{"p":"nano-gpt","id":"xiaomi/mimo-v2.5-pro","n":"MiMo V2.5 Pro","d":"MiMo pro model for strong multimodal reasoning and agent execution","ci":0.435,"co":0.87,"cr":0.0036,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"nano-gpt","id":"xiaomi/mimo-v2-flash-thinking","n":"MiMo V2 Flash (Thinking)","f":"mimo","d":"MiMo flash model for fast multimodal assistance and agent workflows","ci":0.102,"co":0.306,"ctx":256000,"out":32768,"mi":"t","mo":"t","rs":1,"rd":"2025-12-17","lu":"2025-12-17"},{"p":"nano-gpt","id":"anthropic/claude-haiku-latest","n":"Claude Haiku Latest","d":"Fast Claude model for responsive assistance, classification, and lightweight agents","ci":1,"co":5,"cr":0.1,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-29","lu":"2026-03-29"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.7:thinking","n":"Claude 4.7 Opus Thinking","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"cr":0.4998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.6:thinking:max","n":"Claude 4.6 Opus Thinking Max","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.6:thinking:low","n":"Claude 4.6 Opus Thinking Low","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"nano-gpt","id":"anthropic/claude-sonnet-4.6:thinking","n":"Claude Sonnet 4.6 Thinking","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.992,"co":14.993999999999998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.7","n":"Claude 4.7 Opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"cr":0.4998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-04-16","lu":"2026-04-16"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.6:thinking:medium","n":"Claude 4.6 Opus Thinking Medium","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.8:thinking","n":"Claude Opus 4.8 Thinking","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"cr":0.4998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"nano-gpt","id":"anthropic/claude-sonnet-latest","n":"Claude Sonnet Latest","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.992,"co":14.994,"cr":0.2992,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-01","lu":"2026-03-01"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.8","n":"Claude Opus 4.8","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"cr":0.4998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"nano-gpt","id":"anthropic/claude-opus-latest","n":"Claude Opus Latest","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"cr":0.4998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-03-29","lu":"2026-03-29"},{"p":"nano-gpt","id":"anthropic/claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":2.992,"co":14.993999999999998,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-02-17"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.6:thinking","n":"Claude 4.6 Opus Thinking","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"nano-gpt","id":"anthropic/claude-opus-4.6","n":"Claude 4.6 Opus","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":4.998,"co":25.007,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-02-05","lu":"2026-02-05"},{"p":"nano-gpt","id":"tencent/Hunyuan-MT-7B","n":"Hunyuan MT 7B","f":"hunyuan","d":"Translation model for multilingual conversion, localization, and cross-language workflows","ci":10,"co":20,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2025-09-18","lu":"2025-09-18"},{"p":"nano-gpt","id":"tencent/hy3-preview","n":"Tencent: Hy3 preview","d":"Tencent Hy reasoning model for coding, instruction following, and agent tasks","ci":0.066,"co":0.26,"cr":0.029,"ctx":262144,"out":262144,"mi":"t","mo":"t","rd":"2026-04-23","lu":"2026-04-23"},{"p":"nano-gpt","id":"chutesai/Mistral-Small-3.2-24B-Instruct-2506","n":"Mistral Small 3.2 24b Instruct","f":"chutesai","d":"Efficient Mistral model for fast chat, extraction, and production assistants","ci":0.2,"co":0.4,"ctx":128000,"out":131072,"mi":"t","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"dmind/dmind-1-mini","n":"DMind-1-Mini","f":"gpt","d":"Compact GPT model for low-latency assistance and high-volume workloads","ci":0.2,"co":0.4,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2025-06-01","lu":"2025-06-01"},{"p":"nano-gpt","id":"dmind/dmind-1","n":"DMind-1","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":0.3,"co":0.6,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2025-06-01","lu":"2025-06-01"},{"p":"nano-gpt","id":"MarinaraSpaghetti/NemoMix-Unleashed-12B","n":"NemoMix 12B Unleashed","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"deepcogito/cogito-v1-preview-qwen-32B","n":"Cogito v1 Preview Qwen 32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":1.7999999999999998,"co":1.7999999999999998,"ctx":128000,"out":32768,"mi":"t","mo":"t","rd":"2025-05-10","lu":"2025-05-10"},{"p":"nano-gpt","id":"cohere/command-r","n":"Cohere: Command R","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":0.476,"co":1.428,"ctx":128000,"out":4096,"mi":"t","mo":"t","rd":"2024-03-11","lu":"2024-03-11"},{"p":"nano-gpt","id":"cohere/command-r-plus-08-2024","n":"Cohere: Command R+","f":"command-r","d":"Cohere retrieval model for long-context chat and enterprise RAG workflows","ci":2.856,"co":14.246,"ctx":128000,"out":4096,"mi":"t","mo":"t","tc":1,"rd":"2024-08-30","lu":"2024-08-30"},{"p":"nano-gpt","id":"stepfun/step-3.7-flash:thinking","n":"Step 3.7 Flash Thinking","d":"StepFun flash model for efficient multimodal reasoning, coding, and tool use","ci":0.2,"co":1.15,"cr":0.04,"ctx":256000,"out":256000,"mi":"tiv","mo":"t","rs":1,"tc":1,"rd":"2026-05-29","lu":"2026-05-29"},{"p":"nano-gpt","id":"nex-agi/deepseek-v3.1-nex-n1","n":"DeepSeek V3.1 Nex N1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27999999999999997,"co":0.42000000000000004,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2025-12-10","lu":"2025-12-10"},{"p":"nano-gpt","id":"undi95/remm-slerp-l2-13b","n":"ReMM SLERP 13B","f":"llama","d":"Open Llama multimodal model for image understanding and text reasoning","ci":0.7989999999999999,"co":1.2069999999999999,"ctx":6144,"out":4096,"mi":"tp","mo":"t","rd":"2025-01-01","lu":"2025-01-01"},{"p":"nano-gpt","id":"nothingiisreal/L3.1-70B-Celeste-V0.1-BF16","n":"Llama 3.1 70B Celeste v0.1","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-07-23","lu":"2024-07-23"},{"p":"nano-gpt","id":"zai-org/glm-4.7-flash-original:thinking","n":"GLM 4.7 Flash Original Thinking","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"nano-gpt","id":"zai-org/glm-4.7","n":"GLM 4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.15,"co":0.8,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-29","lu":"2026-01-29"},{"p":"nano-gpt","id":"zai-org/GLM-4.5-Air:thinking","n":"GLM 4.5 Air (Thinking)","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.12,"co":0.8,"ctx":128000,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"zai-org/GLM-4.5:thinking","n":"GLM 4.5 (Thinking)","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.3,"co":1.3,"ctx":128000,"out":65536,"mi":"t","mo":"t","rs":1,"rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"zai-org/glm-4.5","n":"GLM 4.5","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.3,"co":1.3,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"zai-org/glm-5-original","n":"GLM 5 Original","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"nano-gpt","id":"zai-org/glm-5.1","n":"GLM 5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.3,"co":2.55,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"nano-gpt","id":"zai-org/glm-4.7-original","n":"GLM 4.7 Original","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.11,"ctx":200000,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-22","lu":"2025-12-22"},{"p":"nano-gpt","id":"zai-org/glm-4.7-original:thinking","n":"GLM 4.7 Original Thinking","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.6,"co":2.2,"cr":0.11,"ctx":200000,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-22","lu":"2025-12-22"},{"p":"nano-gpt","id":"zai-org/glm-4.7-flash-original","n":"GLM 4.7 Flash Original","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"nano-gpt","id":"zai-org/glm-4.7:thinking","n":"GLM 4.7 Thinking","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.2,"co":0.8,"ctx":200000,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-22","lu":"2025-12-22"},{"p":"nano-gpt","id":"zai-org/glm-5.1:thinking","n":"GLM 5.1 Thinking","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.3,"co":2.55,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"nano-gpt","id":"zai-org/glm-5:thinking","n":"GLM 5 Thinking","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.3,"co":2.55,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"nano-gpt","id":"zai-org/glm-4.6v","n":"GLM 4.6V","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"ctx":128000,"out":24000,"mi":"ti","mo":"t","rd":"2025-12-11","lu":"2025-12-11"},{"p":"nano-gpt","id":"zai-org/glm-4.6-original","n":"GLM 4.6 Original","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.35,"co":1.4,"ctx":256000,"out":65535,"mi":"t","mo":"t","rs":1,"rd":"2025-12-11","lu":"2025-12-11"},{"p":"nano-gpt","id":"zai-org/glm-5-original:thinking","n":"GLM 5 Original Thinking","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.2,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"nano-gpt","id":"zai-org/glm-4.7-flash:thinking","n":"GLM 4.7 Flash Thinking","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"nano-gpt","id":"zai-org/glm-4.6v-flash-original","n":"GLM 4.6V Flash","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.1,"co":0.4,"ctx":128000,"out":24000,"mi":"tiv","mo":"t","rd":"2025-12-08","lu":"2025-12-08"},{"p":"nano-gpt","id":"zai-org/glm-4.6v-original","n":"GLM 4.6V Original","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.6,"co":0.9,"ctx":128000,"out":24000,"mi":"ti","mo":"t","rd":"2025-12-08","lu":"2025-12-08"},{"p":"nano-gpt","id":"zai-org/glm-latest","n":"GLM Latest","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.75,"co":2.6,"cr":0.15,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-03","lu":"2026-05-03"},{"p":"nano-gpt","id":"zai-org/glm-4.7-flash","n":"GLM 4.7 Flash","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.07,"co":0.4,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-01-19","lu":"2026-01-19"},{"p":"nano-gpt","id":"zai-org/GLM-4.6-turbo:thinking","n":"GLM 4.6 Turbo (Thinking)","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":1,"co":3,"ctx":200000,"out":204800,"mi":"t","mo":"t","rs":1,"rd":"2025-10-02","lu":"2025-10-02"},{"p":"nano-gpt","id":"zai-org/GLM-4.5-Air","n":"GLM 4.5 Air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.12,"co":0.8,"ctx":128000,"out":98304,"mi":"t","mo":"t","tc":1,"rd":"2025-04-15","lu":"2025-04-15"},{"p":"nano-gpt","id":"zai-org/GLM-4.6-turbo","n":"GLM 4.6 Turbo","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":1,"co":3,"ctx":200000,"out":204800,"mi":"t","mo":"t","rd":"2025-10-02","lu":"2025-10-02"},{"p":"nano-gpt","id":"zai-org/glm-5","n":"GLM 5","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.3,"co":2.55,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-11","lu":"2026-02-11"},{"p":"nano-gpt","id":"Infermatic/MN-12B-Inferor-v0.0","n":"Mistral Nemo Inferor 12B","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.25499999999999995,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"shisa-ai/shisa-v2.1-llama3.3-70b","n":"Shisa V2.1 Llama 3.3 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.5,"co":0.5,"ctx":32768,"out":4096,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"shisa-ai/shisa-v2-llama3.3-70b","n":"Shisa V2 Llama 3.3 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.5,"co":0.5,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"abacusai/Dracarys-72B-Instruct","n":"Llama 3.1 70B Dracarys 2","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2025-08-02","lu":"2025-08-02"},{"p":"nano-gpt","id":"deepseek-ai/DeepSeek-V3.1:thinking","n":"DeepSeek V3.1 Thinking","f":"deepseek-thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2,"co":0.7,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-08-21","lu":"2025-08-21"},{"p":"nano-gpt","id":"deepseek-ai/DeepSeek-V3.1-Terminus","n":"DeepSeek V3.1 Terminus","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.25,"co":0.7,"ctx":128000,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-08-02","lu":"2025-08-02"},{"p":"nano-gpt","id":"deepseek-ai/deepseek-v3.2-exp-thinking","n":"DeepSeek V3.2 Exp Thinking","f":"deepseek-thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27999999999999997,"co":0.42000000000000004,"ctx":163840,"out":65536,"mi":"t","mo":"t","rs":1,"rd":"2025-09-29","lu":"2025-09-29"},{"p":"nano-gpt","id":"deepseek-ai/DeepSeek-R1-0528","n":"DeepSeek R1 0528","f":"deepseek","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.4,"co":1.7,"ctx":128000,"out":163840,"mi":"t","mo":"t","rs":1,"rd":"2025-05-28","lu":"2025-05-28"},{"p":"nano-gpt","id":"deepseek-ai/DeepSeek-V3.1","n":"DeepSeek V3.1","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.2,"co":0.7,"ctx":128000,"out":65536,"mi":"tp","mo":"t","rd":"2025-07-26","lu":"2025-07-26"},{"p":"nano-gpt","id":"deepseek-ai/deepseek-v3.2-exp","n":"DeepSeek V3.2 Exp","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27999999999999997,"co":0.42000000000000004,"ctx":163840,"out":65536,"mi":"t","mo":"t","rd":"2025-09-29","lu":"2025-09-29"},{"p":"nano-gpt","id":"deepseek-ai/DeepSeek-V3.1-Terminus:thinking","n":"DeepSeek V3.1 Terminus (Thinking)","f":"deepseek-thinking","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.25,"co":0.7,"ctx":128000,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-09-22","lu":"2025-09-22"},{"p":"nano-gpt","id":"arcee-ai/trinity-large-thinking","n":"Trinity Large Thinking","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":0.9,"ctx":262144,"out":80000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"nano-gpt","id":"arcee-ai/trinity-mini","n":"Trinity Mini","f":"trinity-mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.045000000000000005,"co":0.15,"ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-12-01","lu":"2025-12-01"},{"p":"nano-gpt","id":"soob3123/GrayLine-Qwen3-8B","n":"Grayline Qwen3 8B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.3,"co":0.3,"ctx":16384,"out":32768,"mi":"t","mo":"t","rd":"2025-09-25","lu":"2025-09-25"},{"p":"nano-gpt","id":"soob3123/Veiled-Calla-12B","n":"Veiled Calla 12B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.3,"co":0.3,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2025-04-13","lu":"2025-04-13"},{"p":"nano-gpt","id":"soob3123/amoral-gemma3-27B-v2","n":"Amoral Gemma3 27B v2","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.3,"co":0.3,"ctx":32768,"out":8192,"mi":"t","mo":"t","rd":"2025-05-23","lu":"2025-05-23"},{"p":"nano-gpt","id":"meganova-ai/manta-mini-1.0","n":"Manta Mini 1.0","f":"nova","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.02,"co":0.16,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2025-12-20","lu":"2025-12-20"},{"p":"nano-gpt","id":"meganova-ai/manta-flash-1.0","n":"Manta Flash 1.0","f":"nova","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.02,"co":0.16,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2025-12-20","lu":"2025-12-20"},{"p":"nano-gpt","id":"meganova-ai/manta-pro-1.0","n":"Manta Pro 1.0","f":"nova","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.060000000000000005,"co":0.5,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-12-20","lu":"2025-12-20"},{"p":"nano-gpt","id":"qwen/Qwen3.6-35B-A3B:thinking","n":"Qwen3.6 35B A3B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.112,"co":0.8,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-19","lu":"2026-04-19"},{"p":"nano-gpt","id":"qwen/qwen3-coder-plus","n":"Qwen3 Coder Plus","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":1,"co":5,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-09-17","lu":"2025-09-17"},{"p":"nano-gpt","id":"qwen/Qwen3.6-35B-A3B","n":"Qwen3.6 35B A3B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.112,"co":0.8,"ctx":262144,"out":16384,"mi":"tiv","mo":"t","rd":"2026-04-17","lu":"2026-04-17"},{"p":"nano-gpt","id":"qwen/qwen3-32b","n":"Qwen 3 32b","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.1,"co":0.3,"ctx":41000,"out":32768,"mi":"tp","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen/Qwen3-VL-235B-A22B-Instruct","n":"Qwen3 VL 235B A22B Instruct","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":1.2,"ctx":128000,"out":262144,"mi":"ti","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen/qwen3-max","n":"Qwen3 Max","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":1.08018,"co":5.4009,"ctx":256000,"out":32768,"mi":"t","mo":"t","rd":"2025-09-05","lu":"2025-09-05"},{"p":"nano-gpt","id":"qwen/qwen3.5-397b-a17b-thinking","n":"Qwen3.5 397B A17B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":258048,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-02-16","lu":"2026-02-16"},{"p":"nano-gpt","id":"qwen/Qwen3-Next-80B-A3B-Instruct","n":"Qwen3 Next 80B A3B (Instruct)","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":0.65,"ctx":256000,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2025-09-11","lu":"2025-09-11"},{"p":"nano-gpt","id":"qwen/Qwen3-235B-A22B-Instruct-2507-TEE","n":"Qwen 3 235b A22B 2507 (TEE)","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.13,"co":0.5,"ctx":256000,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2025-07-25","lu":"2025-07-25"},{"p":"nano-gpt","id":"qwen/qwq-32b-preview","n":"Qwen QwQ 32B Preview","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.2,"co":0.2,"ctx":32768,"out":32768,"mi":"t","mo":"t","rd":"2025-02-27","lu":"2025-02-27"},{"p":"nano-gpt","id":"qwen/qwen3-next-80b-a3b-thinking","n":"Qwen3 Next 80B A3B (Thinking)","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.15,"co":0.65,"ctx":256000,"out":32768,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen/qwen3-coder","n":"Qwen 3 Coder 480B","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.13,"co":0.5,"ctx":262000,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2026-03-17","lu":"2026-03-17"},{"p":"nano-gpt","id":"qwen/Qwen3-235B-A22B-Thinking-2507","n":"Qwen 3 235b A22B 2507 Thinking","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.3,"co":0.5,"ctx":256000,"out":262144,"mi":"t","mo":"t","rd":"2025-09-11","lu":"2025-09-11"},{"p":"nano-gpt","id":"qwen/qwen3.5-plus","n":"Qwen3.5 Plus","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":2.4,"cr":0.04,"ctx":983616,"out":65536,"mi":"tiv","mo":"t","rd":"2026-02-16","lu":"2026-02-16"},{"p":"nano-gpt","id":"qwen/qwen3.5-plus-thinking","n":"Qwen3.5 Plus Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.4,"co":2.4,"cr":0.04,"ctx":983616,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-02-16","lu":"2026-02-16"},{"p":"nano-gpt","id":"qwen/qwen3-235b-a22b","n":"Qwen 3 235b A22B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":0.5,"ctx":41000,"out":32768,"mi":"tp","mo":"t","tc":1,"rd":"2025-04-29","lu":"2025-04-29"},{"p":"nano-gpt","id":"qwen/qwen-2.5-72b-instruct","n":"Qwen2.5 72B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.357,"co":0.408,"ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-07-03","lu":"2025-07-03"},{"p":"nano-gpt","id":"qwen/qwen3-coder-next","n":"Qwen3 Coder Next","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.15,"co":1.5,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"rd":"2025-12-08","lu":"2025-12-08"},{"p":"nano-gpt","id":"qwen/qwen3.5-9b","n":"Qwen3.5 9B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.05,"co":0.15,"ctx":256000,"out":65536,"mi":"ti","mo":"t","rs":1,"rd":"2026-03-10","lu":"2026-03-10"},{"p":"nano-gpt","id":"qwen/qwen3.5-397b-a17b","n":"Qwen3.5 397B A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.6,"co":3.6,"ctx":258048,"out":65536,"mi":"tiv","mo":"t","ow":1,"rd":"2026-02-16","lu":"2026-02-16"},{"p":"nano-gpt","id":"qwen/qwen3-coder-flash","n":"Qwen3 Coder Flash","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.3,"co":1.5,"ctx":128000,"out":65536,"mi":"t","mo":"t","rd":"2025-09-17","lu":"2025-09-17"},{"p":"nano-gpt","id":"qwen/Qwen2.5-Coder-32B-Instruct","n":"Qwen 2.5 Coder 32b","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.2006,"co":0.2006,"ctx":32000,"out":8192,"mi":"t","mo":"t","rd":"2025-07-03","lu":"2025-07-03"},{"p":"nano-gpt","id":"qwen/Qwen3-8B","n":"Qwen 3 8B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.47,"co":0.47,"ctx":41000,"out":32768,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen/qwen3-14b","n":"Qwen 3 14b","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.08,"co":0.24,"ctx":41000,"out":32768,"mi":"t","mo":"t","rd":"2024-01-01","lu":"2024-01-01"},{"p":"nano-gpt","id":"qwen/Qwen3-235B-A22B-Instruct-2507","n":"Qwen 3 235b A22B 2507","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.13,"co":0.5,"ctx":256000,"out":262144,"mi":"t","mo":"t","tc":1,"rd":"2025-07-25","lu":"2025-07-25"},{"p":"nano-gpt","id":"qwen/qwen3-30b-a3b","n":"Qwen3 30B A3B","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.1,"co":0.3,"ctx":41000,"out":32768,"mi":"t","mo":"t","rd":"2025-02-27","lu":"2025-02-27"},{"p":"nano-gpt","id":"amazon/nova-lite-v1","n":"Amazon Nova Lite 1.0","f":"nova-lite","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.0595,"co":0.238,"ctx":300000,"out":5120,"mi":"t","mo":"t","rd":"2024-12-03","lu":"2024-12-03"},{"p":"nano-gpt","id":"amazon/nova-pro-v1","n":"Amazon Nova Pro 1.0","f":"nova-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.7989999999999999,"co":3.1959999999999997,"ctx":300000,"out":32000,"mi":"t","mo":"t","rd":"2024-12-03","lu":"2024-12-03"},{"p":"nano-gpt","id":"amazon/nova-micro-v1","n":"Amazon Nova Micro 1.0","f":"nova-micro","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.0357,"co":0.1394,"ctx":128000,"out":5120,"mi":"t","mo":"t","rd":"2024-12-03","lu":"2024-12-03"},{"p":"nano-gpt","id":"amazon/nova-2-lite-v1","n":"Amazon Nova 2 Lite","f":"nova","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.5099999999999999,"co":4.25,"ctx":1000000,"out":65535,"mi":"t","mo":"t","rd":"2024-12-03","lu":"2024-12-03"},{"p":"nano-gpt","id":"alibaba/qwen3.6-flash","n":"Qwen3.6 Flash","f":"qwen3.6","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.19,"co":1.16,"ctx":991800,"out":65536,"mi":"tiv","mo":"t","rd":"2026-04-17","lu":"2026-04-17"},{"p":"nano-gpt","id":"alibaba/qwen3.6-27b","n":"Qwen3.6 27B","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.203,"co":2.24,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rd":"2026-04-23","lu":"2026-04-23"},{"p":"nano-gpt","id":"alibaba/qwen3.6-27b:thinking","n":"Qwen3.6 27B Thinking","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.203,"co":2.24,"ctx":260096,"out":65536,"mi":"tiv","mo":"t","rs":1,"rd":"2026-04-23","lu":"2026-04-23"},{"p":"nano-gpt","id":"aion-labs/aion-rp-llama-3.1-8b","n":"Llama 3.1 8b (uncensored)","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.2006,"co":0.2006,"ctx":32768,"out":16384,"mi":"t","mo":"t","rd":"2024-07-23","lu":"2024-07-23"},{"p":"nano-gpt","id":"aion-labs/aion-2.5","n":"AionLabs: Aion-2.5","d":"General-purpose chat model for instruction following, writing, and analysis","ci":1,"co":3,"cr":0.35,"ctx":131072,"out":32768,"mi":"t","mo":"t","rd":"2026-03-20","lu":"2026-03-20"},{"p":"nano-gpt","id":"aion-labs/aion-1.0-mini","n":"Aion 1.0 mini (DeepSeek)","f":"deepseek","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.7989999999999999,"co":1.394,"ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-02-20","lu":"2025-02-20"},{"p":"nano-gpt","id":"aion-labs/aion-2.0","n":"AionLabs: Aion-2.0","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.8,"co":1.6,"ctx":131072,"out":32768,"mi":"t","mo":"t","rd":"2026-02-23","lu":"2026-02-23"},{"p":"nano-gpt","id":"aion-labs/aion-1.0","n":"Aion 1.0","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":3.995,"co":7.99,"ctx":65536,"out":8192,"mi":"t","mo":"t","rd":"2025-02-01","lu":"2025-02-01"},{"p":"nano-gpt","id":"pamanseau/OpenReasoning-Nemotron-32B","n":"OpenReasoning Nemotron 32B","f":"nemotron","d":"Nemotron model for efficient reasoning, coding, and specialized AI agents","ci":0.1,"co":0.4,"ctx":32768,"out":65536,"mi":"t","mo":"t","rs":1,"rd":"2025-08-21","lu":"2025-08-21"},{"p":"nano-gpt","id":"LatitudeGames/Wayfarer-Large-70B-Llama-3.3","n":"Llama 3.3 70B Wayfarer","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.700000007,"co":0.700000007,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2025-02-20","lu":"2025-02-20"},{"p":"nano-gpt","id":"baseten/Kimi-K2-Instruct-FP4","n":"Kimi K2 0711 Instruct FP4","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.1,"co":2,"ctx":128000,"out":131072,"mi":"t","mo":"t","rd":"2025-07-11","lu":"2025-07-11"},{"p":"nano-gpt","id":"inflection/inflection-3-pi","n":"Inflection 3 Pi","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.499,"co":9.996,"ctx":8000,"out":4096,"mi":"t","mo":"t","rd":"2024-10-11","lu":"2024-10-11"},{"p":"nano-gpt","id":"inflection/inflection-3-productivity","n":"Inflection 3 Productivity","f":"gpt","d":"GPT model for general reasoning, writing, coding, and tool-assisted tasks","ci":2.499,"co":9.996,"ctx":8000,"out":4096,"mi":"t","mo":"t","rd":"2024-10-11","lu":"2024-10-11"},{"p":"nano-gpt","id":"ReadyArt/MS3.2-The-Omega-Directive-24B-Unslop-v2.0","n":"Omega Directive 24B Unslop v2.0","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.5,"co":0.5,"ctx":16384,"out":32768,"mi":"t","mo":"t","rd":"2025-12-08","lu":"2025-12-08"},{"p":"nano-gpt","id":"MiniMaxAI/MiniMax-M1-80k","n":"MiniMax M1 80K","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.6052,"co":2.4225000000000003,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rd":"2025-06-16","lu":"2025-06-16"},{"p":"nano-gpt","id":"upstage/solar-pro-3","n":"Solar Pro 3","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.15,"co":0.6,"cr":0.015,"ctx":128000,"out":128000,"mi":"t","mo":"t","rd":"2026-03-03","lu":"2026-03-03"},{"p":"nano-gpt","id":"allenai/olmo-3-32b-think","n":"Olmo 3 32B Think","f":"allenai","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.3,"co":0.44999999999999996,"ctx":128000,"out":8192,"mi":"t","mo":"t","rs":1,"rd":"2025-11-01","lu":"2025-11-01"},{"p":"nano-gpt","id":"essentialai/rnj-1-instruct","n":"RNJ-1 Instruct 8B","f":"rnj","d":"General-purpose chat model for instruction following, writing, and analysis","ci":0.15,"co":0.15,"ctx":128000,"out":8192,"mi":"t","mo":"t","rd":"2025-12-13","lu":"2025-12-13"},{"p":"nano-gpt","id":"deepseek/deepseek-v4-flash","n":"DeepSeek V4 Flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"cr":0.028,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"nano-gpt","id":"deepseek/deepseek-v4-flash:thinking","n":"DeepSeek V4 Flash (Thinking)","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"cr":0.028,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"nano-gpt","id":"deepseek/deepseek-v4-pro-cheaper:thinking","n":"DeepSeek V4 Pro Cheaper (Thinking)","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-25","lu":"2026-04-25"},{"p":"nano-gpt","id":"deepseek/deepseek-prover-v2-671b","n":"DeepSeek Prover v2 671B","f":"deepseek","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1,"co":2.5,"ctx":160000,"out":16384,"mi":"t","mo":"t","rd":"2025-04-30","lu":"2025-04-30"},{"p":"nano-gpt","id":"deepseek/deepseek-latest","n":"DeepSeek Latest","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":1.1,"co":2.2,"cr":0.11,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-03","lu":"2026-05-03"},{"p":"nano-gpt","id":"deepseek/deepseek-v4-pro","n":"DeepSeek V4 Pro","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.1,"co":2.2,"cr":0.11,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"nano-gpt","id":"deepseek/deepseek-v4-pro:thinking","n":"DeepSeek V4 Pro (Thinking)","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.1,"co":2.2,"cr":0.11,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-24","lu":"2026-04-24"},{"p":"nano-gpt","id":"deepseek/deepseek-v3.2-speciale","n":"DeepSeek V3.2 Speciale","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27999999999999997,"co":0.42000000000000004,"ctx":163000,"out":65536,"mi":"tp","mo":"t","rs":1,"rd":"2025-12-02","lu":"2025-12-02"},{"p":"nano-gpt","id":"deepseek/deepseek-v4-pro-cheaper","n":"DeepSeek V4 Pro Cheaper","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1048576,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-25","lu":"2026-04-25"},{"p":"nano-gpt","id":"deepseek/deepseek-v3.2:thinking","n":"DeepSeek V3.2 Thinking","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27999999999999997,"co":0.42000000000000004,"ctx":163000,"out":65536,"mi":"tp","mo":"t","rs":1,"tc":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"nano-gpt","id":"deepseek/deepseek-v3.2","n":"DeepSeek V3.2","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.27999999999999997,"co":0.42000000000000004,"ctx":163000,"out":65536,"mi":"tp","mo":"t","tc":1,"rd":"2025-12-01","lu":"2025-12-01"},{"p":"nano-gpt","id":"minimax/minimax-m3:thinking","n":"MiniMax M3 Thinking","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":512000,"out":80000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"nano-gpt","id":"minimax/minimax-m2.5","n":"MiniMax M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"nano-gpt","id":"minimax/minimax-m2.1","n":"MiniMax M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.33,"co":1.32,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2025-12-19","lu":"2025-12-19"},{"p":"nano-gpt","id":"minimax/minimax-m2.7-turbo","n":"MiniMax M2.7 Turbo","d":"Efficient MiniMax model for quick assistance, coding, and routine automation","ci":0.6,"co":2.4,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"nano-gpt","id":"minimax/minimax-01","n":"MiniMax 01","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.1394,"co":1.1219999999999999,"ctx":1000192,"out":16384,"mi":"tp","mo":"t","rd":"2025-01-15","lu":"2025-01-15"},{"p":"nano-gpt","id":"minimax/minimax-m3","n":"MiniMax M3","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":512000,"out":80000,"mi":"ti","mo":"t","tc":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"nano-gpt","id":"minimax/minimax-m2-her","n":"MiniMax M2-her","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.30200000000000005,"co":1.2069999999999999,"ctx":65532,"out":2048,"mi":"t","mo":"t","rd":"2026-01-24","lu":"2026-01-24"},{"p":"nano-gpt","id":"minimax/minimax-latest","n":"MiniMax Latest","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0.3,"co":1.2,"cr":0.06,"ctx":512000,"out":80000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-03","lu":"2026-05-03"},{"p":"nano-gpt","id":"minimax/minimax-m2.7","n":"MiniMax M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"nano-gpt","id":"kwaipilot/kat-coder-pro-v2","n":"KAT Coder Pro V2","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ci":0.3,"co":1.2,"ctx":256000,"out":80000,"mi":"t","mo":"t","rd":"2026-03-28","lu":"2026-03-28"},{"p":"nano-gpt","id":"mlabonne/NeuralDaredevil-8B-abliterated","n":"Neural Daredevil 8B abliterated","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.44,"co":0.44,"ctx":8192,"out":8192,"mi":"t","mo":"t","rd":"2024-12-01","lu":"2024-12-01"},{"p":"nano-gpt","id":"VongolaChouko/Starcannon-Unleashed-12B-v1.0","n":"Mistral Nemo Starcannon 12b v1","f":"mistral-nemo","d":"Mistral model for multilingual chat, reasoning, and tool-assisted workflows","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-07-01","lu":"2024-07-01"},{"p":"nano-gpt","id":"TEE/minimax-m2.5","n":"MiniMax M2.5 TEE","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.2,"co":1.38,"ctx":196608,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"nano-gpt","id":"TEE/glm-4.7","n":"GLM 4.7 TEE","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0.85,"co":3.3,"ctx":131000,"out":65535,"mi":"t","mo":"t","rd":"2026-01-29","lu":"2026-01-29"},{"p":"nano-gpt","id":"TEE/llama3-3-70b","n":"Llama 3.3 70B","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":2,"co":2,"ctx":128000,"out":16384,"mi":"t","mo":"t","rd":"2025-07-03","lu":"2025-07-03"},{"p":"nano-gpt","id":"TEE/gemma-4-31b-it","n":"Gemma 4 31B IT TEE","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.15,"co":0.46,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-26","lu":"2026-05-26"},{"p":"nano-gpt","id":"TEE/qwen3-30b-a3b-instruct-2507","n":"Qwen3 30B A3B Instruct 2507 TEE","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.15,"co":0.44999999999999996,"ctx":262000,"out":32768,"mi":"t","mo":"t","rd":"2025-07-29","lu":"2025-07-29"},{"p":"nano-gpt","id":"TEE/glm-5.1","n":"GLM 5.1 TEE","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.5,"co":5.25,"cr":0.3,"ctx":202752,"out":65535,"mi":"t","mo":"t","rd":"2026-04-20","lu":"2026-04-20"},{"p":"nano-gpt","id":"TEE/deepseek-v4-pro","n":"DeepSeek V4 Pro TEE","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.5,"co":5.25,"cr":0.15,"ctx":800000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-25","lu":"2026-04-25"},{"p":"nano-gpt","id":"TEE/deepseek-v4-pro:thinking","n":"DeepSeek V4 Pro Thinking TEE","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.5,"co":5.25,"cr":0.15,"ctx":800000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-29","lu":"2026-04-29"},{"p":"nano-gpt","id":"TEE/gemma4-31b:thinking","n":"Gemma 4 31B Thinking TEE","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.45,"co":1,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"rd":"2026-05-02","lu":"2026-05-02"},{"p":"nano-gpt","id":"TEE/kimi-k2.5","n":"Kimi K2.5 TEE","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0.3,"co":1.9,"ctx":128000,"out":65535,"mi":"t","mo":"t","rd":"2026-01-29","lu":"2026-01-29"},{"p":"nano-gpt","id":"TEE/qwen2.5-vl-72b-instruct","n":"Qwen2.5 VL 72B TEE","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.7,"co":0.7,"ctx":65536,"out":8192,"mi":"ti","mo":"t","rd":"2025-02-01","lu":"2025-02-01"},{"p":"nano-gpt","id":"TEE/gpt-oss-120b","n":"GPT-OSS 120B TEE","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":2,"co":2,"ctx":131072,"out":16384,"mi":"t","mo":"t","rd":"2025-08-05","lu":"2025-08-05"},{"p":"nano-gpt","id":"TEE/qwen3.5-27b","n":"Qwen3.5 27B TEE","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.3,"co":2.4,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rd":"2026-03-13","lu":"2026-03-13"},{"p":"nano-gpt","id":"TEE/gemma-4-26b-a4b-uncensored","n":"Gemma 4 26B A4B Uncensored TEE","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.15,"co":0.7,"ctx":65536,"out":65536,"mi":"ti","mo":"t","tc":1,"rd":"2026-05-23","lu":"2026-05-23"},{"p":"nano-gpt","id":"TEE/deepseek-v3.1","n":"DeepSeek V3.1 TEE","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":1,"co":2.5,"ctx":164000,"out":8192,"mi":"t","mo":"t","rd":"2025-08-21","lu":"2025-08-21"},{"p":"nano-gpt","id":"TEE/kimi-k2.6","n":"Kimi K2.6 TEE","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":1.5,"co":5.25,"cr":0.375,"ctx":262144,"out":65536,"mi":"ti","mo":"t","tc":1,"rd":"2026-04-21","lu":"2026-04-21"},{"p":"nano-gpt","id":"TEE/qwen3.5-397b-a17b","n":"Qwen3.5 397B A17B TEE","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.6,"co":3.6,"ctx":258048,"out":65536,"mi":"t","mo":"t","rd":"2026-02-28","lu":"2026-02-28"},{"p":"nano-gpt","id":"TEE/kimi-k2.5-thinking","n":"Kimi K2.5 Thinking TEE","f":"kimi-thinking","d":"Kimi reasoning model for long-horizon research, planning, and tool use","ci":0.3,"co":1.9,"ctx":128000,"out":65535,"mi":"t","mo":"t","rs":1,"rd":"2026-01-29","lu":"2026-01-29"},{"p":"nano-gpt","id":"TEE/qwen3.6-35b-a3b-uncensored","n":"Qwen3.6 35B A3B Uncensored TEE","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.3,"co":1.5,"ctx":131072,"out":131072,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-23","lu":"2026-05-23"},{"p":"nano-gpt","id":"TEE/glm-4.7-flash","n":"GLM 4.7 Flash TEE","f":"glm-flash","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0.15,"co":0.5,"ctx":203000,"out":65535,"mi":"t","mo":"t","rd":"2026-01-19","lu":"2026-01-19"},{"p":"nano-gpt","id":"TEE/gemma4-31b","n":"Gemma 4 31B","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.45,"co":1,"ctx":262144,"out":131072,"mi":"t","mo":"t","rd":"2026-04-04","lu":"2026-04-04"},{"p":"nano-gpt","id":"TEE/gemma-3-27b-it","n":"Gemma 3 27B TEE","f":"gemma","d":"Open Gemma instruction model for efficient chat and self-hosted deployments","ci":0.2,"co":0.8,"ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-03-10","lu":"2025-03-10"},{"p":"nano-gpt","id":"TEE/gpt-oss-20b","n":"GPT-OSS 20B TEE","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.2,"co":0.8,"ctx":131072,"out":8192,"mi":"t","mo":"t","rd":"2025-08-05","lu":"2025-08-05"},{"p":"nano-gpt","id":"TEE/glm-5","n":"GLM 5 TEE","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.2,"co":3.5,"ctx":203000,"out":65535,"mi":"t","mo":"t","rd":"2026-02-11","lu":"2026-02-11"},{"p":"nano-gpt","id":"TEE/glm-5.1-thinking","n":"GLM 5.1 Thinking TEE","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1.5,"co":5.25,"cr":0.3,"ctx":202752,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-04-20","lu":"2026-04-20"},{"p":"nano-gpt","id":"TEE/deepseek-v3.2","n":"DeepSeek V3.2 TEE","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.5,"co":1,"ctx":164000,"out":65536,"mi":"t","mo":"t","rd":"2025-12-01","lu":"2025-12-01"},{"p":"nano-gpt","id":"TEE/qwen3.5-122b-a10b","n":"Qwen3.5 122B A10B TEE","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ci":0.46,"co":3.68,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-26","lu":"2026-05-26"},{"p":"nano-gpt","id":"Sao10K/L3.1-70B-Hanami-x1","n":"Llama 3.1 70B Hanami","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-07-23","lu":"2024-07-23"},{"p":"nano-gpt","id":"Sao10K/L3-8B-Stheno-v3.2","n":"Sao10K Stheno 8b","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.2006,"co":0.2006,"ctx":16384,"out":8192,"mi":"t","mo":"t","rd":"2024-11-29","lu":"2024-11-29"},{"p":"nano-gpt","id":"Sao10K/L3.3-70B-Euryale-v2.3","n":"Llama 3.3 70B Euryale","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":20480,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"Sao10K/L3.1-70B-Euryale-v2.2","n":"Llama 3.1 70B Euryale","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.306,"co":0.357,"ctx":20480,"out":16384,"mi":"t","mo":"t","rd":"2024-07-23","lu":"2024-07-23"},{"p":"nano-gpt","id":"Steelskull/L3.3-MS-Evalebis-70b","n":"MS Evalebis 70b","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"Steelskull/L3.3-MS-Nevoria-70b","n":"Steelskull Nevoria 70b","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"Steelskull/L3.3-Cu-Mai-R1-70b","n":"Llama 3.3 70B Cu Mai","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"Steelskull/L3.3-MS-Evayale-70B","n":"Evayale 70b ","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"Steelskull/L3.3-Nevoria-R1-70b","n":"Steelskull Nevoria R1 70b","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.49299999999999994,"co":0.49299999999999994,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"nano-gpt","id":"Steelskull/L3.3-Electra-R1-70b","n":"Steelskull Electra R1 70b","f":"llama","d":"Open Llama instruction model for multilingual chat, reasoning, and coding","ci":0.69989,"co":0.69989,"ctx":16384,"out":16384,"mi":"t","mo":"t","rd":"2024-12-06","lu":"2024-12-06"},{"p":"moark","id":"MiniMax-M2.1","n":"MiniMax-M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":2.1,"co":8.4,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"moark","id":"GLM-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":3.5,"co":14,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"lilac","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.7,"co":3.5,"cr":0.2,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"lilac","id":"minimaxai/minimax-m3","n":"MiniMax M3","f":"minimax-m3","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.28,"co":1.1,"cr":0.05,"ctx":1048576,"out":1048576,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-01"},{"p":"lilac","id":"google/gemma-4-31b-it","n":"Gemma 4 31B IT","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.11,"co":0.35,"ctx":262100,"out":262100,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-02","lu":"2026-04-02"},{"p":"lilac","id":"zai-org/glm-5.2","n":"GLM 5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0.9,"co":3,"cr":0.27,"ctx":524288,"out":524288,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"ambient","id":"moonshotai/kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.75,"co":3.5,"cr":0.16,"cw":0,"ctx":262144,"out":262144,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"ambient","id":"moonshotai/kimi-k2.6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"cr":0.2,"cw":0,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"ambient","id":"zai-org/GLM-5.1-FP8","n":"GLM 5.1","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":1.4,"co":4.4,"cr":0,"cw":0,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"neon","id":"gemini-3-flash","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"neon","id":"claude-sonnet-4","n":"Claude Sonnet 4.5","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"neon","id":"claude-opus-4-5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"neon","id":"gpt-5","n":"GPT-5","f":"gpt","d":"Original GPT-5 workhorse for reasoning, coding, writing, and tool workflows","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"neon","id":"gpt-5-1","n":"GPT-5.1","f":"gpt","d":"Sharper GPT-5 generation for coding, product work, and tool-assisted tasks","ci":1.25,"co":10,"cr":0.125,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-09-30","rd":"2025-11-13","lu":"2025-11-13"},{"p":"neon","id":"gemini-3-pro","n":"Gemini 3 Pro Preview","f":"gemini-pro","d":"Preview Gemini flagship for complex reasoning, coding, and rich multimodal prompts","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-11-18","lu":"2025-11-18"},{"p":"neon","id":"gpt-5-2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"neon","id":"gemini-3-1-flash-lite","n":"Gemini 3.1 Flash Lite Preview","f":"gemini-flash-lite","d":"Low-latency Gemini model for high-volume multimodal and agent workloads","ci":0.25,"co":1.5,"cr":0.025,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-03-03","lu":"2026-03-03"},{"p":"neon","id":"claude-sonnet-4-5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"neon","id":"claude-opus-4-7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"neon","id":"gemini-2-5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"neon","id":"gpt-5-4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"neon","id":"claude-opus-4-1","n":"Claude Opus 4.1 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":15,"co":75,"cr":1.5,"cw":18.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-08-05","lu":"2025-08-05"},{"p":"neon","id":"gpt-5-4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"neon","id":"gemini-2-5-flash","n":"Gemini 2.5 Flash","f":"gemini-flash","d":"Fast Gemini workhorse for multimodal apps where latency and price matter","ci":0.3,"co":2.5,"cr":0.03,"ctx":1048576,"out":65536,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"neon","id":"gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open GPT reasoning model for self-hosted agents and controllable deployments","ci":0.072,"co":0.28,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"neon","id":"gpt-5-5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"neon","id":"gemini-3-1-pro","n":"Gemini 3.1 Pro Preview Custom Tools","f":"gemini-pro","d":"Advanced Gemini model for complex reasoning, coding, and multimodal analysis","ci":2,"co":12,"cr":0.2,"ctx":1048576,"out":65536,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"neon","id":"claude-haiku-4-5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"neon","id":"claude-opus-4-6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"neon","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"neon","id":"gpt-5-nano","n":"GPT-5 Nano","f":"gpt-nano","d":"Tiny GPT-5 lane for routing, extraction, classification, and bulk jobs","ci":0.05,"co":0.4,"cr":0.005,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"neon","id":"claude-sonnet-4-6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":1000000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"neon","id":"gpt-5-4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":1050000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"neon","id":"gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.05,"co":0.2,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"upstage","id":"solar-pro2","n":"solar-pro2","f":"solar-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":0.25,"ctx":65536,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2025-05-20","lu":"2025-05-20"},{"p":"upstage","id":"solar-pro3","n":"solar-pro3","f":"solar-pro","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":0.25,"co":0.25,"ctx":131072,"out":8192,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-03","rd":"2026-01","lu":"2026-01"},{"p":"upstage","id":"solar-mini","n":"solar-mini","f":"solar-mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.15,"co":0.15,"ctx":32768,"out":4096,"mi":"t","mo":"t","tc":1,"kn":"2024-09","rd":"2024-06-12","lu":"2025-04-22"},{"p":"zhipuai-coding-plan","id":"glm-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-27","lu":"2026-03-27"},{"p":"zhipuai-coding-plan","id":"glm-5v-turbo","n":"GLM-5V-Turbo","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"tivp","mo":"t","rs":1,"tc":1,"rd":"2026-04-01","lu":"2026-04-01"},{"p":"zhipuai-coding-plan","id":"glm-5-turbo","n":"GLM-5-Turbo","f":"glm","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":200000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-03-16","lu":"2026-03-16"},{"p":"zhipuai-coding-plan","id":"glm-4.5-air","n":"GLM-4.5-Air","f":"glm-air","d":"Efficient GLM model for fast reasoning, coding, and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":131072,"out":98304,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-28","lu":"2025-07-28"},{"p":"zhipuai-coding-plan","id":"glm-4.6v","n":"GLM-4.6V","f":"glm","d":"GLM vision model for visual reasoning, documents, and multimodal agents","ci":0.3,"co":0.9,"ctx":128000,"out":32768,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-08","lu":"2025-12-08"},{"p":"zhipuai-coding-plan","id":"glm-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"zhipuai-coding-plan","id":"glm-4.7","n":"GLM-4.7","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-12-22","lu":"2025-12-22"},{"p":"chutes","id":"moonshotai/Kimi-K2.6-TEE","n":"Kimi K2.6 TEE","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.66,"co":3.5,"cr":0.33,"ctx":262144,"out":65535,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2026-04-21","lu":"2026-04-21"},{"p":"chutes","id":"moonshotai/Kimi-K2.5-TEE","n":"Kimi K2.5 TEE","f":"kimi-k2","d":"Earlier Kimi frontier model for long-context agents, coding, and multimodal work","ci":0.44,"co":2,"cr":0.22,"ctx":262144,"out":65535,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2026-01","lu":"2026-01"},{"p":"chutes","id":"google/gemma-4-31B-turbo-TEE","n":"gemma 4 31B turbo TEE","f":"gemma","d":"Largest Gemma 4 instruction model for open, self-hosted chat and reasoning","ci":0.12,"co":0.37,"cr":0.06,"ctx":131072,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-02","lu":"2026-04-02"},{"p":"chutes","id":"Qwen/Qwen3-32B-TEE","n":"Qwen3 32B TEE","f":"qwen","d":"Dense open Qwen model for self-hosted chat, reasoning, and coding","ci":0.104,"co":0.416,"cr":0.052,"ctx":40960,"out":40960,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-04","lu":"2025-04"},{"p":"chutes","id":"Qwen/Qwen3.6-27B-TEE","n":"Qwen3.6 27B TEE","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.3,"co":2,"cr":0.15,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"chutes","id":"Qwen/Qwen3-235B-A22B-Thinking-2507-TEE","n":"Qwen3 235B A22B Thinking 2507 TEE","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.2989,"co":1.1957,"cr":0.14945,"ctx":262144,"out":262144,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07","lu":"2026-06-21"},{"p":"chutes","id":"Qwen/Qwen3.5-397B-A17B-TEE","n":"Qwen3.5 397B A17B TEE","f":"qwen","d":"Large open Qwen multimodal MoE for visual agents and long technical tasks","ci":0.45,"co":3,"cr":0.225,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-15","lu":"2026-02-15"},{"p":"chutes","id":"unsloth/Mistral-Nemo-Instruct-2407-TEE","n":"Mistral Nemo Instruct 2407 TEE","f":"mistral-nemo","d":"Efficient Mistral-NVIDIA open model for multilingual chat and local deployment","ci":0.0245,"co":0.0978,"cr":0.01225,"ctx":131072,"out":131072,"mi":"t","mo":"t","ow":1,"kn":"2024-07","rd":"2024-07-01","lu":"2024-07-01"},{"p":"chutes","id":"zai-org/GLM-5-TEE","n":"GLM 5 TEE","f":"glm","d":"General GLM flagship for coding, analysis, and tool-heavy engineering workflows","ci":0.95,"co":2.55,"cr":0.475,"ctx":202752,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"chutes","id":"zai-org/GLM-5.1-TEE","n":"GLM 5.1 TEE","f":"glm","d":"Strong GLM coding model for agentic engineering, terminals, and repository generation","ci":0.98,"co":3.08,"cr":0.49,"ctx":202752,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-07","lu":"2026-04-07"},{"p":"chutes","id":"zai-org/GLM-5.2-TEE","n":"GLM 5.2 TEE","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":0.7,"ctx":1048576,"out":65535,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"chutes","id":"deepseek-ai/DeepSeek-V3.2-TEE","n":"DeepSeek V3.2 TEE","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":1,"co":1,"cr":0.5,"ctx":131072,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12","lu":"2026-06-21"},{"p":"chutes","id":"MiniMaxAI/MiniMax-M2.5-TEE","n":"MiniMax M2.5 TEE","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.15,"co":1.2,"cr":0.075,"ctx":196608,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M2.1","n":"MiniMax-M2.1","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M2","n":"MiniMax-M2","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"ctx":196608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal coding model for long-context reasoning and agent tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":1000000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-25"},{"p":"minimax-cn-coding-plan","id":"MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0,"co":0,"cr":0,"cw":0,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"deepseek","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek V4 lane for economical reasoning, coding, and long-context work","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"deepseek","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Open MoE flagship with million-token context for coding and long agent runs","ci":0.435,"co":0.87,"cr":0.003625,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-04-24"},{"p":"deepseek","id":"deepseek-reasoner","n":"DeepSeek Reasoner","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"deepseek","id":"deepseek-chat","n":"DeepSeek Chat","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0.14,"co":0.28,"cr":0.0028,"ctx":1000000,"out":384000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-09","rd":"2025-12-01","lu":"2026-02-28"},{"p":"wafer.ai","id":"deepseek-v4-flash","n":"DeepSeek V4 Flash","f":"deepseek-flash","d":"Fast DeepSeek model for efficient chat, coding help, and agent loops","ci":0.14,"co":0.28,"cr":0.01,"cw":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-05-30"},{"p":"wafer.ai","id":"Qwen3.6-35B-A3B","n":"Qwen3.6-35B-A3B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.15,"co":1,"cr":0.02,"cw":0,"ctx":256000,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-05-11","lu":"2026-05-30"},{"p":"wafer.ai","id":"qwen3.7-max","n":"Qwen3.7-Max","f":"qwen3.7-max","d":"Flagship model for demanding analysis, coding, and production agent workflows","ci":5,"co":15,"cr":0.5,"cw":0,"ctx":256000,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-21","lu":"2026-05-30"},{"p":"wafer.ai","id":"deepseek-v4-pro","n":"DeepSeek V4 Pro","f":"deepseek-thinking","d":"Flagship DeepSeek model for coding, reasoning, and agentic work","ci":1.74,"co":3.48,"cr":0.02,"cw":0,"ctx":1000000,"out":384000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-05","rd":"2026-04-24","lu":"2026-05-30"},{"p":"wafer.ai","id":"Qwen3.5-397B-A17B","n":"Qwen3.5-397B-A17B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.43,"co":2.6,"cr":0.04,"cw":0,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-02-16","lu":"2026-06-01"},{"p":"wafer.ai","id":"Kimi-K2.6","n":"Kimi-K2.6","f":"kimi-k2","d":"Kimi multimodal agent model for visual understanding, coding, and planning","ci":0.68,"co":3.15,"cr":0.07,"cw":0,"ctx":262144,"out":65536,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-05-13","lu":"2026-06-01"},{"p":"wafer.ai","id":"GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.2,"co":4.1,"cr":0.2,"cw":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-22"},{"p":"wafer.ai","id":"GLM-5.1","n":"GLM-5.1","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":1,"co":3.2,"cr":0.1,"cw":0,"ctx":202752,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-04-07","lu":"2026-06-01"},{"p":"minimax","id":"MiniMax-M2.1","n":"MiniMax-M2.1","f":"minimax","d":"Earlier MiniMax agent model for practical coding and productivity tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12-23","lu":"2025-12-23"},{"p":"minimax","id":"MiniMax-M2.5-highspeed","n":"MiniMax-M2.5-highspeed","f":"minimax","d":"High-speed MiniMax model for low-latency coding and agent workflows","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-13","lu":"2026-02-13"},{"p":"minimax","id":"MiniMax-M2.7-highspeed","n":"MiniMax-M2.7-highspeed","f":"minimax","d":"Low-latency M2.7 variant for interactive coding plans and agent loops","ci":0.6,"co":2.4,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"minimax","id":"MiniMax-M2","n":"MiniMax-M2","f":"minimax","d":"Efficient open MiniMax model built for coding agents and tool-heavy workflows","ci":0.3,"co":1.2,"ctx":196608,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-10-27","lu":"2025-10-27"},{"p":"minimax","id":"MiniMax-M2.5","n":"MiniMax-M2.5","f":"minimax","d":"Prior MiniMax coding model for agent workflows, office edits, and automation","ci":0.3,"co":1.2,"cr":0.03,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-12"},{"p":"minimax","id":"MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.3,"co":1.2,"cr":0.06,"ctx":1000000,"out":128000,"mi":"tiv","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-01","lu":"2026-06-25"},{"p":"minimax","id":"MiniMax-M2.7","n":"MiniMax-M2.7","f":"minimax","d":"Open MiniMax flagship for coding agents, office automation, and complex environments","ci":0.3,"co":1.2,"cr":0.06,"cw":0.375,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"github-copilot","id":"claude-sonnet-4.5","n":"Claude Sonnet 4.5 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-07-31","rd":"2025-09-29","lu":"2025-09-29"},{"p":"github-copilot","id":"claude-sonnet-4","n":"Claude Sonnet 4 (latest)","f":"claude-sonnet","d":"Balanced Claude model for coding, analysis, agent workflows, and cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":216000,"out":16000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-05-22","lu":"2025-05-22"},{"p":"github-copilot","id":"gemini-2.5-pro","n":"Gemini 2.5 Pro","f":"gemini-pro","d":"Google's proven reasoning model for coding, math, and multimodal analysis","ci":1.25,"co":10,"cr":0.125,"ctx":128000,"out":64000,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-06-17","lu":"2025-06-17"},{"p":"github-copilot","id":"claude-haiku-4.5","n":"Claude Haiku 4.5 (latest)","f":"claude-haiku","d":"Fast Claude lane for lightweight agents, office tasks, and responsive chat","ci":1,"co":5,"cr":0.1,"cw":1.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-02-28","rd":"2025-10-15","lu":"2025-10-15"},{"p":"github-copilot","id":"gemini-3.5-flash","n":"Gemini 3.5 Flash","f":"gemini-flash","d":"Fast Gemini model balancing multimodal reasoning, tool use, and cost","ci":1.5,"co":9,"cr":0.15,"ctx":200000,"out":64000,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-05-19","lu":"2026-05-19"},{"p":"github-copilot","id":"kimi-k2.7-code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.19,"ctx":256000,"out":32000,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"github-copilot","id":"claude-sonnet-5","n":"Claude Sonnet 5","f":"claude-sonnet","d":"Everyday Claude agent model for coding, planning, browsing, and general work","ci":2,"co":10,"cr":0.2,"cw":2.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-30","lu":"2026-06-30"},{"p":"github-copilot","id":"gpt-5.4-nano","n":"GPT-5.4 nano","f":"gpt-nano","d":"Cheapest GPT-5.4 lane for simple routing, extraction, and bulk automation","ci":0.2,"co":1.25,"cr":0.02,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"github-copilot","id":"claude-opus-4.7","n":"Claude Opus 4.7","f":"claude-opus","d":"Stronger Opus tier for advanced software work and high-stakes reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-04-16","lu":"2026-04-16"},{"p":"github-copilot","id":"mai-code-1-flash-picker","n":"MAI-Code-1-Flash","f":"mai","d":"Microsoft coding model built for fast, efficient assistance in everyday developer workflows","ci":0.75,"co":4.5,"cr":0.075,"ctx":256000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2025-12","rd":"2026-06-02","lu":"2026-06-08"},{"p":"github-copilot","id":"gpt-5.2","n":"GPT-5.2","f":"gpt","d":"Reliable GPT generation for broad coding, writing, and tool-assisted product work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"github-copilot","id":"gpt-5.3-codex","n":"GPT-5.3 Codex","f":"gpt-codex","d":"Coding-optimized GPT model for repository edits, reviews, and agentic software work","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-05","lu":"2026-02-05"},{"p":"github-copilot","id":"claude-opus-4.8","n":"Claude Opus 4.8","f":"claude-opus","d":"Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":64000,"mi":"tip","mo":"t","rs":1,"tc":1,"rd":"2026-05-28","lu":"2026-05-28"},{"p":"github-copilot","id":"claude-fable-5","n":"Claude Fable 5","f":"claude-fable","d":"Claude model for creative writing, analysis, and controlled agent workflows","ci":10,"co":50,"cr":1,"cw":12.5,"ctx":1000000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2026-01-31","rd":"2026-06-09","lu":"2026-06-09"},{"p":"github-copilot","id":"claude-opus-4.5","n":"Claude Opus 4.5 (latest)","f":"claude-opus","d":"Flagship Claude model for deep reasoning, coding, and long-horizon agents","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-03-31","rd":"2025-11-24","lu":"2025-11-24"},{"p":"github-copilot","id":"gpt-5.4","n":"GPT-5.4","f":"gpt","d":"Agent-ready GPT for coding and computer-use workflows at a lower cost","ci":2.5,"co":15,"cr":0.25,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-05","lu":"2026-03-05"},{"p":"github-copilot","id":"gpt-5.4-mini","n":"GPT-5.4 mini","f":"gpt-mini","d":"Strong small GPT for coding subagents, quick tool use, and high-volume work","ci":0.75,"co":4.5,"cr":0.075,"ctx":400000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-03-17","lu":"2026-03-17"},{"p":"github-copilot","id":"gpt-4.1","n":"GPT-4.1","f":"gpt","d":"Long-lived GPT workhorse for coding, instruction following, and production apps","ci":2,"co":8,"cr":0.5,"ctx":128000,"out":16384,"mi":"tip","mo":"t","tc":1,"kn":"2024-04","rd":"2025-04-14","lu":"2025-04-14"},{"p":"github-copilot","id":"gemini-3.1-pro-preview","n":"Gemini 3.1 Pro Preview","f":"gemini-pro","d":"Reasoning-first Gemini preview for agentic coding and complex problem solving","ci":2,"co":12,"cr":0.2,"ctx":200000,"out":64000,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2026-02-19","lu":"2026-02-19"},{"p":"github-copilot","id":"claude-sonnet-4.6","n":"Claude Sonnet 4.6","f":"claude-sonnet","d":"Claude workhorse for coding agents, careful analysis, and production cost control","ci":3,"co":15,"cr":0.3,"cw":3.75,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2026-02-17","lu":"2026-03-13"},{"p":"github-copilot","id":"gpt-5-mini","n":"GPT-5 Mini","f":"gpt-mini","d":"Small GPT-5 for responsive agents, coding help, and everyday automation","ci":0.25,"co":2,"cr":0.025,"ctx":264000,"out":64000,"mi":"ti","mo":"t","rs":1,"tc":1,"kn":"2024-05-30","rd":"2025-08-07","lu":"2025-08-07"},{"p":"github-copilot","id":"gemini-3-flash-preview","n":"Gemini 3 Flash Preview","f":"gemini-flash","d":"New Gemini flash lane bringing frontier-style multimodal reasoning to cheaper runs","ci":0.5,"co":3,"cr":0.05,"ctx":128000,"out":64000,"mi":"tivap","mo":"t","rs":1,"tc":1,"kn":"2025-01","rd":"2025-12-17","lu":"2025-12-17"},{"p":"github-copilot","id":"claude-opus-4.6","n":"Claude Opus 4.6","f":"claude-opus","d":"High-end Claude for difficult coding, planning, and slower expert reasoning","ci":5,"co":25,"cr":0.5,"cw":6.25,"ctx":200000,"out":32000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-05-31","rd":"2026-02-05","lu":"2026-03-13"},{"p":"github-copilot","id":"gpt-5.2-codex","n":"GPT-5.2 Codex","f":"gpt-codex","d":"Code-specialist GPT for repository edits, reviews, and long-running software agents","ci":1.75,"co":14,"cr":0.175,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-08-31","rd":"2025-12-11","lu":"2025-12-11"},{"p":"github-copilot","id":"gpt-5.5","n":"GPT-5.5","f":"gpt","d":"Default frontier GPT for coding, computer use, research, and knowledge work","ci":5,"co":30,"cr":0.5,"ctx":400000,"out":128000,"mi":"tip","mo":"t","rs":1,"tc":1,"kn":"2025-12-01","rd":"2026-04-23","lu":"2026-04-23"},{"p":"clarifai","id":"moonshotai/chat-completion/models/Kimi-K2_6","n":"Kimi K2.6","f":"kimi-k2","d":"Multimodal Kimi workhorse for agent loops, coding tasks, and visual context","ci":0.95,"co":4,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-04-21","lu":"2026-04-21"},{"p":"clarifai","id":"minimaxai/chat-completion/models/MiniMax-M2_5-high-throughput","n":"MiniMax-M2.5 High Throughput","f":"minimax","d":"MiniMax model for chat, coding, office work, and agentic tasks","ci":0.3,"co":1.2,"ctx":204800,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-02-12","lu":"2026-02-25"},{"p":"clarifai","id":"openai/chat-completion/models/gpt-oss-120b-high-throughput","n":"GPT OSS 120B High Throughput","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.09,"co":0.36,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2026-02-25"},{"p":"clarifai","id":"openai/chat-completion/models/gpt-oss-20b","n":"GPT OSS 20B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.045,"co":0.18,"ctx":131072,"out":16384,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-12-12"},{"p":"clarifai","id":"mistralai/completion/models/Ministral-3-14B-Reasoning-2512","n":"Ministral 3 14B Reasoning 2512","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":2.5,"co":1.7,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-12","rd":"2025-12-01","lu":"2025-12-12"},{"p":"clarifai","id":"mistralai/completion/models/Ministral-3-3B-Reasoning-2512","n":"Ministral 3 3B Reasoning 2512","f":"ministral","d":"Compact Mistral model for edge, latency-sensitive, and cost-efficient workloads","ci":1.039,"co":0.54825,"ctx":262144,"out":262144,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-12","lu":"2026-02-25"},{"p":"clarifai","id":"deepseek-ai/deepseek-ocr/models/DeepSeek-OCR","n":"DeepSeek OCR","f":"deepseek","d":"OCR model for extracting structured text from documents and screenshots","ci":0.2,"co":0.7,"ctx":8192,"out":8192,"mi":"ti","mo":"t","ow":1,"rd":"2025-10-20","lu":"2026-02-25"},{"p":"clarifai","id":"qwen/qwenLM/models/Qwen3-30B-A3B-Thinking-2507","n":"Qwen3 30B A3B Thinking 2507","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0.36,"co":1.3,"ctx":262144,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-07-31","lu":"2026-02-25"},{"p":"clarifai","id":"qwen/qwenLM/models/Qwen3-30B-A3B-Instruct-2507","n":"Qwen3 30B A3B Instruct 2507","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0.3,"co":0.5,"ctx":262144,"out":262144,"mi":"t","mo":"t","tc":1,"ow":1,"rd":"2025-07-30","lu":"2026-02-25"},{"p":"clarifai","id":"qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-Instruct","n":"Qwen3 Coder 30B A3B Instruct","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0.11458,"co":0.74812,"ctx":262144,"out":65536,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-31","lu":"2026-02-12"},{"p":"clarifai","id":"arcee_ai/AFM/models/trinity-mini","n":"Trinity Mini","f":"trinity-mini","d":"Efficient model for low-latency assistance, extraction, and routine automation","ci":0.045,"co":0.15,"ctx":131072,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2025-12","lu":"2026-02-25"},{"p":"clarifai","id":"clarifai/main/models/mm-poly-8b","n":"MM Poly 8B","f":"mm-poly","d":"Multimodal model for analyzing text, images, documents, and rich media","ci":0.658,"co":1.11,"ctx":32768,"out":4096,"mi":"tiv","mo":"t","rd":"2025-06","lu":"2026-02-25"},{"p":"the-grid-ai","id":"agent-prime","n":"Agent Prime","d":"Preview model for early access evaluation, prototyping, and compatibility testing","ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-04","lu":"2026-05-19"},{"p":"the-grid-ai","id":"agent-max","n":"Agent Max","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-04","lu":"2026-05-19"},{"p":"the-grid-ai","id":"text-standard","n":"Text Standard","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":128000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-26","lu":"2026-05-19"},{"p":"the-grid-ai","id":"code-prime","n":"Code Prime","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":128000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-04","lu":"2026-05-19"},{"p":"the-grid-ai","id":"text-prime","n":"Text Prime","d":"Reasoning model for deliberate analysis, multi-step problem solving, and tool use","ctx":128000,"out":30000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-02-26","lu":"2026-05-19"},{"p":"the-grid-ai","id":"code-max","n":"Code Max","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-05-04","lu":"2026-05-19"},{"p":"the-grid-ai","id":"agent-standard","n":"Agent Standard","d":"Preview model for early access evaluation, prototyping, and compatibility testing","ctx":128000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-04","lu":"2026-05-19"},{"p":"the-grid-ai","id":"text-max","n":"Text Max","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ctx":1000000,"out":128000,"mi":"ti","mo":"t","rs":1,"tc":1,"rd":"2026-03-24","lu":"2026-05-19"},{"p":"the-grid-ai","id":"code-standard","n":"Code Standard","d":"Coding model for repository understanding, refactors, and agentic engineering tasks","ctx":128000,"out":16000,"mi":"t","mo":"t","rs":1,"tc":1,"rd":"2026-05-04","lu":"2026-05-19"},{"p":"synthetic","id":"hf:moonshotai/Kimi-K2.7-Code","n":"Kimi K2.7 Code","f":"kimi-k2","d":"Coding-focused Kimi model, stronger on long-horizon repo work with less overthinking","ci":0.95,"co":4,"cr":0.95,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-01","rd":"2026-06-12","lu":"2026-06-12"},{"p":"synthetic","id":"hf:zai-org/GLM-4.7-Flash","n":"GLM-4.7-Flash","f":"glm-flash","d":"Budget GLM lane for fast coding help, routing, and everyday automation","ci":0.1,"co":0.5,"cr":0.1,"ctx":196608,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2026-01-19","lu":"2026-01-19"},{"p":"synthetic","id":"hf:zai-org/GLM-5.2","n":"GLM-5.2","f":"glm","d":"Open flagship GLM for long-horizon coding agents and million-token context work","ci":1.4,"co":4.4,"cr":1.4,"ctx":524288,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-13","lu":"2026-06-13"},{"p":"synthetic","id":"hf:MiniMaxAI/MiniMax-M3","n":"MiniMax-M3","f":"minimax","d":"MiniMax multimodal model for long-context coding, perception, and agent planning","ci":0.6,"co":1.2,"cr":0.6,"ctx":524288,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-06-12","lu":"2026-06-12"},{"p":"synthetic","id":"hf:openai/gpt-oss-120b","n":"GPT OSS 120B","f":"gpt-oss","d":"Open-weight GPT model for self-hosted reasoning and instruction-following workloads","ci":0.1,"co":0.1,"cr":0.1,"ctx":131072,"out":32768,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2025-08-05","lu":"2025-08-05"},{"p":"synthetic","id":"hf:Qwen/Qwen3.6-27B","n":"Qwen3.6 27B","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0.45,"co":3.6,"cr":0.45,"ctx":262144,"out":65536,"mi":"ti","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"synthetic","id":"hf:nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4","n":"Nemotron 3 Super 120B A12B","f":"nemotron","d":"Nemotron middle tier for collaborative agents and high-volume reasoning workloads","ci":0.3,"co":1,"cr":0.3,"ctx":262144,"out":65536,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"rd":"2026-03-11","lu":"2026-03-11"},{"p":"iflowcn","id":"qwen3-coder-plus","n":"Qwen3-Coder-Plus","f":"qwen","d":"Qwen coding model for software agents, repository edits, and code reasoning","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-01","lu":"2025-07-01"},{"p":"iflowcn","id":"deepseek-v3","n":"DeepSeek-V3","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0,"co":0,"ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-12-26","lu":"2024-12-26"},{"p":"iflowcn","id":"kimi-k2","n":"Kimi-K2","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0,"co":0,"ctx":128000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2024-10","rd":"2024-12-01","lu":"2024-12-01"},{"p":"iflowcn","id":"qwen3-32b","n":"Qwen3-32B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":128000,"out":32000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-10","rd":"2024-12-01","lu":"2024-12-01"},{"p":"iflowcn","id":"qwen3-max-preview","n":"Qwen3-Max-Preview","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0,"co":0,"ctx":256000,"out":32000,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"iflowcn","id":"qwen3-max","n":"Qwen3-Max","f":"qwen","d":"Flagship Qwen model for complex reasoning, coding, and agentic workflows","ci":0,"co":0,"ctx":256000,"out":32000,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"iflowcn","id":"qwen3-235b","n":"Qwen3-235B-A22B","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-10","rd":"2024-12-01","lu":"2024-12-01"},{"p":"iflowcn","id":"glm-4.6","n":"GLM-4.6","f":"glm","d":"Flagship GLM model for hybrid reasoning, coding, and agentic engineering","ci":0,"co":0,"ctx":200000,"out":128000,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-10","rd":"2024-12-01","lu":"2025-11-13"},{"p":"iflowcn","id":"qwen3-235b-a22b-thinking-2507","n":"Qwen3-235B-A22B-Thinking","f":"qwen","d":"Qwen reasoning model for deliberate problem solving, math, and coding","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-01","lu":"2025-07-01"},{"p":"iflowcn","id":"deepseek-r1","n":"DeepSeek-R1","f":"deepseek-thinking","d":"DeepSeek reasoning model for multi-step analysis, math, coding, and tools","ci":0,"co":0,"ctx":128000,"out":32000,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-20","lu":"2025-01-20"},{"p":"iflowcn","id":"qwen3-vl-plus","n":"Qwen3-VL-Plus","f":"qwen","d":"Qwen vision-language model for visual reasoning, documents, and agent tasks","ci":0,"co":0,"ctx":256000,"out":32000,"mi":"ti","mo":"t","tc":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"iflowcn","id":"qwen3-235b-a22b-instruct","n":"Qwen3-235B-A22B-Instruct","f":"qwen","d":"Qwen instruction model for multilingual chat, reasoning, and tool use","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2025-04","rd":"2025-07-01","lu":"2025-07-01"},{"p":"iflowcn","id":"kimi-k2-0905","n":"Kimi-K2-0905","f":"kimi-k2","d":"Kimi model for long-context chat, coding, and agentic reasoning","ci":0,"co":0,"ctx":256000,"out":64000,"mi":"t","mo":"t","tc":1,"kn":"2024-12","rd":"2025-09-05","lu":"2025-09-05"},{"p":"iflowcn","id":"deepseek-v3.2","n":"DeepSeek-V3.2-Exp","f":"deepseek","d":"DeepSeek chat model for instruction following, coding, and analysis","ci":0,"co":0,"ctx":128000,"out":64000,"mi":"t","mo":"t","tc":1,"ow":1,"kn":"2024-12","rd":"2025-01-01","lu":"2025-01-01"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2.5-tts","n":"MiMo-V2.5-TTS","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2.5-pro","n":"MiMo-V2.5-Pro","f":"mimo","d":"Stronger MiMo Pro tier for multimodal reasoning and coding-agent execution","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2-pro","n":"MiMo-V2-Pro","f":"mimo","d":"Earlier MiMo Pro model for multimodal agents, reasoning, and code tasks","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"t","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2-tts","n":"MiMo-V2-TTS","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2-omni","n":"MiMo-V2-Omni","f":"mimo","d":"MiMo omni model for text, image, video, audio, and agents","ci":0,"co":0,"cr":0,"ctx":262144,"out":131072,"mi":"tiavp","mo":"t","rs":1,"tc":1,"kn":"2024-12","rd":"2026-03-18","lu":"2026-03-18"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2.5","n":"MiMo-V2.5","f":"mimo","d":"Open MiMo model for multimodal coding agents and long-context automation","ci":0,"co":0,"cr":0,"ctx":1048576,"out":131072,"mi":"tiav","mo":"t","rs":1,"tc":1,"ow":1,"kn":"2024-12","rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2.5-tts-voicedesign","n":"MiMo-V2.5-TTS-VoiceDesign","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"xiaomi-token-plan-sgp","id":"mimo-v2.5-tts-voiceclone","n":"MiMo-V2.5-TTS-VoiceClone","f":"mimo","d":"Speech generation model for controllable voice, narration, and audio delivery","ci":0,"co":0,"ctx":8192,"out":8192,"mi":"t","mo":"a","ow":1,"rd":"2026-04-22","lu":"2026-04-22"},{"p":"claudinio","id":"claudinio","n":"Claudinio","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":0.5,"co":2,"cr":0.15,"ctx":256000,"out":64000,"mi":"tiav","mo":"t","rs":1,"tc":1,"kn":"2026-05","rd":"2026-05-12","lu":"2026-06-02"},{"p":"claudinio","id":"claudius","n":"Claudius","d":"Multimodal reasoning model for visual analysis, planning, and tool use","ci":3,"co":8,"cr":0.9,"ctx":256000,"out":64000,"mi":"tiav","mo":"t","rs":1,"tc":1,"kn":"2026-05","rd":"2026-05-12","lu":"2026-05-12"}]}