Qwen3-Coder: Agentic Coding in the World
Qwen ● Covered by 2 sources
Alibaba released Qwen3-Coder, a code-focused AI model available in multiple sizes, with the flagship variant containing 480 billion parameters and 35 billion active parameters supporting context lengths up to 256,000 tokens natively and 1 million with extrapolation. The model achieved state-of-the-art performance among open models on agentic coding, browser-use, and tool-use benchmarks, with performance comparable to Claude Sonnet 4. The release enables developers to access an advanced open-source coding model for agentic tasks that was previously only available through commercial alternatives.
Why it matters
GITHUB HUGGING FACE MODELSCOPE DISCORD Today, we’re announcing Qwen3-Coder, our most agentic code model to date. Qwen3-Coder is available in multiple sizes, but we’re excited to introduce its most powerful variant first: Qwen3-Coder-480B-A35B-Instruct — a 480B-parameter Mixture-of-Experts model with 35B active parameters which supports the context length of 256K tokens natively and 1M tokens with extrapolation methods, offering exceptional performance in both coding and agentic tasks. Qwen3-Coder-480B-A35B-Instruct sets new state-of-the-art results among open models on Agentic Coding, Agentic Browser-Use, and Agentic Tool-Use, comparable to Claude Sonnet 4.