qwen3.6-35b-a3b

Description

Alibaba Qwen3.6-35B-A3B — a 35-billion-parameter native vision-language MoE model with roughly 3 billion active parameters, open-weight, combining linear attention with a sparse mixture of experts for higher inference efficiency. 262,144-token context window, up to 65,536 output tokens, and up to 256 images or 64 videos per request. Accepts text, image, and video input and returns text; supports function calling, built-in tools, and structured outputs.