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.

ProviderContextQuantizationMax outputPriorityState
OpenRouter1M50unknown