ماذا تجيب هذه الصفحة
The Qwen3.5 122B-A10B 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. In terms of
- Alibaba Cloud · Qwen · qwen/qwen3.5-122b-a10b
- text+image+video->text · مسار نموذج الصين
- 262,144 context · 0.26 US$ input