Qwen·released 2025-01-25
Scored on 7 benchmarks. 57% of its scores are independently reproduced — the rest are self-reported or unverified.
Every score, ranked, colored by evidence status.
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| Lech Mazur Writing | 72.9accuracy (%) | unverified | 2025-01-25 |
| MATH level 5 | 67.18accuracy (%) | reproduced· optimized | 2025-01-25 |
| Fiction.LiveBench | 66.7accuracy (%) | unverified | 2025-01-25 |
| GPQA diamond | 41.5accuracy (%) | reproduced· optimized | 2025-01-25 |
| Aider polyglot | 21.8accuracy (%) | unverified· optimized | 2025-01-25 |
| OTIS Mock AIME 2024-2025 | 16.03accuracy (%) | reproduced· optimized | 2025-01-25 |
| FrontierMath-2025-02-28-Private | 1.81accuracy (%) | reproduced· optimized | 2025-01-25 |