Qwen: Qwen3.5-Flash
qwen/qwen3.5-flash-02-23
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...
Modalities
TextImageVideo
In / out price
₩114 / ₩456
1M
Context
1M
Released
Feb 26, 2026
Providers
Several places host the same model. Requests are spread by priority and health, and a local GPU always goes first when one is available.
| Provider | Context | Quantization | Max output | Priority | State |
|---|---|---|---|---|---|
| OpenRouter | 1M | — | — | 50 | unknown |