这页解决什么
Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use.
- 阿里云 · 通义 · qwen/qwen3-coder-30b-a3b-instruct
- text->text · 中国模型路线
- 160,000 context · US$0.07 input