Qwen: Qwen3 235B A22B Thinking 2507
qwen/qwen3-235b-a22b-thinking-2507
Context: 131,072In: $0.23/1MOut: $2.30/1MUpdated 2026-10-04
Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...
Intelligence
12.7
Coding
22.1
Agentic
1.3
Metric profile
Trend over time
Classic evals (curated from model cards)
Alibaba Qwen — manually maintained, may lag releases
MMLU-Pro
76%
GPQA Diamond
69%
MATH-500
96%
Design Arena (ELO by category)
| Category | ELO | Win rate | Rank |
|---|---|---|---|
| website | 1059 | 42% | #107 |
| codecategories | 1039 | 40.8% | #107 |
| 3d | 1013 | 40.4% | #102 |
| gamedev | 970 | 34.2% | #116 |
| dataviz | 957 | 32.6% | #113 |
| uicomponent | 952 | 34% | #113 |