Qwen3-Coder 480B
Alibaba's agentic coding flagship: 480B MoE, 256K context, Apache 2.0.
About
Launched July 22, 2025, Qwen3-Coder-480B-A35B-Instruct is Alibaba's most advanced agentic coding model: 480B total parameters with 35B active per token on a mixture-of-experts architecture, under Apache 2.0. It natively supports 256K context (extrapolatable to 1M tokens) for repo-scale understanding, and set open-source records on agentic coding, browser-use, and tool-use benchmarks — competitive with Claude Sonnet 4 class models. Alibaba paired it with Qwen Code, an open-source terminal coding agent.
Limitations: the full 480B is datacenter-class hardware only; most developers will use it via API or smaller Qwen3-Coder variants. Agentic benchmarks are partly Alibaba-reported — independent replication varies by harness.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Team-reported: SOTA open-source on agentic coding/tool-use benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 44 (models); Trending score 91/100 from researcher assessment.