Qwen: Qwen3.5-Flash
qwen
The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. Compared to the...
- Context window
- 1,000,000 tokens
- Input cost
- $0.07 / 1M
- Output cost
- $0.26 / 1M
- Latency (p50)
- —