这页解决什么
The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It d
- 阿里云 · 通义 · qwen/qwen3.5-397b-a17b
- text+image+video->text · 中国模型路线
- 262,144 context · US$0.39 input