Qwen3-Coder-480B-A35B: open MoE model for agentic coding
Qwen released Qwen3-Coder, a 480B-parameter Mixture-of-Experts model with 35B active parameters, achieving state-of-the-art results among open models on agentic coding, browser-use, and tool-use benchmarks, described as comparable to Claude Sonnet 4.
PUBLISHED2025-07-22
OBSERVED2026-08-11
AGE1y
SOURCES1
- 480B total parameters, 35B active (MoE architecture)
- Native 256K token context, supports up to 1M with extrapolation
- SOTA among open models on Agentic Coding, Agentic Browser-Use, and Agentic Tool-Use
- Available in multiple sizes; the 480B Instruct variant is the first released
COMMUNITY
No curated reactions recorded for this event. Facts and takes are kept in separate layers — community context is added by hand, never blended into the record above.