Qwen3-235B-A22B-2507
Qwen's refreshed 235B MoE flagship - separate Instruct and Thinking models, 256K context, Apache 2.0.
About
Qwen3-235B-A22B-2507 is the July 2025 refresh of Alibaba's 235B-parameter (22B active) mixture-of-experts flagship. The key change from the April 2025 original: Qwen dropped the hybrid 'thinking mode' toggle and instead trains separate Instruct and Thinking variants for best quality in each - Instruct-2507 landed July 21, Thinking-2507 on July 25.
The update adds 256K-token long-context understanding and, with an update in August 2025, support for 1M-token ultra-long inputs. It remains the top open-weight tier of the Qwen3 family across reasoning, coding, and agentic tasks.
Honest limits: running it needs serious hardware - 235B total parameters even at 22B active. Quality comparisons are vendor-reported; the open-weights release is strong but the API-hosted Qwen3-Max variants are a separate tier.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 88; Trending score 84/100 from researcher assessment.
More in the qwen3 family
- Qwen3-32B-AWQ
- Qwen3-Next-80B-A3B
- Qwen3-235B-A22B
- Qwen3-4B-Instruct-2507
- Qwen3-8B
- Qwen3.5
- Qwen3-235B-A22B-Instruct-2507
- Qwen3 235B
- Qwen3-235B-A22B-Thinking-2507
- Qwen3-Omni-Flash
- Qwen3.6-27B
- Qwen3-32B
- Qwen3-Max
- Qwen3.8-Flash-Next
- Qwen3-Coder-Plus
- Qwen3
- Qwen3-0.6B
- Qwen3-4B
- Qwen3-Coder-Next
- Qwen3-4B-2507
- Qwen3-14B
- Qwen3-VL-30B-A3B
- Qwen3-1.7B
- Qwen3-Omni-30B-A3B