The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall...
Details
| Field | Value |
|---|---|
| id | qwen/qwen3.5-35b-a3b |
| context_length | 262144 |
| max_output_tokens | 262144 |
| quantization | unknown |