Qwen3-Coder-30B-A3B

Alibaba / Qwen · Qwen3-Coder
LLMApache-2.0publicTrusted31B params

Coding MoE with 256K context, ~3B active.

Coding MoE with 256K context, ~3B active.

At a glance

  • Task: LLM (InferenceV1)
  • Params: 31B
  • License: Apache-2.0 (commercial use OK)
  • Hardware: 62 GB+ VRAM
  • dtype: Bf16, Fp16