Qwen3-235B-A22B-Instruct (Jul 2025)
#125 of 274 scored.
AlibabaJul 25, 2025Open weights
Capabilities Index
Range 135.7–140.8
138.9
Context
Tokens
—
Input
Per 1M tokens
—
Output
Per 1M tokens
—
Benchmarks
How it scores.
- ARC-AGI-21.3%
64th of 79
Everything else Epoch has for it
- DTBench64.0%
- Aider polyglot59.6%
- Fiction.LiveBench52.9%
- WeirdML38.7%
- LMCA27.7%
- ARC-AGI11.0%
On the index
Around it.
Sources