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

Pricing

Prices are quoted in Korean won. Changes to the exchange rate or margin show up immediately.

Input (prompt)₩114 / 1M
Output (completion)₩456 / 1M
Cache read₩11 / 1M
Cache write

1,000 input tokens plus 1,000 output tokens costs about ₩0.570375.