chutes

Qwen/Qwen3-30B-A3B

Qwen3 30B A3B bf16

Qwen3-30B-A3B is a 30.5B parameter MoE language model (with 3.3B active parameters) that supports switching between thinking and non-thinking modes, making it versatile for complex reasoning tasks like math and coding as well as general dialogue in 100+ languages.