ماذا تجيب هذه الصفحة
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
- Alibaba Cloud · Qwen · qwen/qwen3.5-397b-a17b
- text+image+video->text · مسار نموذج الصين
- 262,144 context · 0.39 US$ input