DeepSeek V4 Flash 0731
by AMD · family deepseek-flash · listed Jul 2026RTW
Official DeepSeek V4 Flash release with enhanced agentic capabilities and integrated DSpark speculative decoding
Current rates USD per 1M tokens
Inputfree
Outputfree
Cache readfree
Cache write—
Against the market
On the record
- Model id
DeepSeek-V4-Flash- Context window
- 1.0M tokens
- Max output
- 384K tokens
- Input modalities
- text
- Output modalities
- text
- Reasoning
- yes
- Tool calls
- yes
- Open weights
- yes
- Knowledge cutoff
- May 2025
- Released
- 2026-07-31
- Last updated
- 2026-07-31
- Provider docs
- developer.amd.com.cn/radeon/tokenfactory
Back of the envelope
≈
Also listed at
- Alibaba Token Planfree in · free out
- Alibaba Token Plan (China)free in · free out
- SCNet Token Planfree in · free out
- Pendrafree in · free out
- SenseNova (China)free in · free out
- Volcengine Ark Coding Planfree in · free out
- InferXfree in · free out
- Adaption$0.0444 in · $0.133 out
- UnoRouter$0.0625 in · $0.125 out
- DevPass (LLM Gateway)$0.076 in · $0.153 out
- LLM Gateway$0.076 in · $0.153 out
- Standard Compute$0.0826 in · $0.165 out
- OpenRouter$0.09 in · $0.18 out
- Kilo Gateway$0.09 in · $0.18 out
- Deep Infra$0.09 in · $0.18 out