Qwen/Qwen3.5-9B
by SiliconFlow (China) · family qwen · listed Mar 2026RTVW
Qwen vision-language model for visual reasoning, documents, and agent tasks
Current rates USD per 1M tokens
Input$0.22
Output$1.74
Cache read—
Cache write—
Against the market
On the record
- Model id
Qwen/Qwen3.5-9B- Context window
- 262K tokens
- Max output
- 66K tokens
- Input modalities
- text, image, video
- Output modalities
- text
- Reasoning
- yes
- Tool calls
- yes
- Open weights
- yes
- Knowledge cutoff
- Apr 2025
- Released
- 2026-03-03
- Last updated
- 2026-03-03
- Provider docs
- cloud.siliconflow.com/models
Back of the envelope
≈
Also listed at
- CrofAI$0.04 in · $0.15 out
- Kilo Gateway$0.05 in · $0.15 out
- NanoGPT$0.05 in · $0.15 out
- Mixlayer$0.10 in · $0.40 out
- LLM Gateway$0.10 in · $0.15 out
- Deep Infra$0.10 in · $0.15 out
- SiliconFlow$0.10 in · $0.15 out
- OpenRouter$0.10 in · $0.15 out
- OVHcloud AI Endpoints$0.12 in · $0.18 out
- Regolo AI$0.15 in · $0.60 out
- Hugging Face$0.17 in · $0.25 out
- Together AI$0.17 in · $0.25 out