| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| MMLU STEM | TaH+ | Accuracy86.6 | 43 | 2mo ago | |
| GPQA | TAC | Pass@1 Accuracy31.9 | 10 | 1mo ago | |
| FrontierSci-olympiad | Score75 | 9 | 24d ago | ||
| FrontierSci research | Score25 | 9 | 24d ago | ||
| Phybench | Score80 | 9 | 24d ago | ||
| BABE | Score58.1 | 9 | 24d ago | ||
| GPQA Diamond | Avg@1690.7 | 8 | 24d ago | ||
| Superchem text-only | Score54.4 | 4 | 24d ago | ||
| MMLU-Pro | DeepSeek V3.2 | Accuracy84.3 | 4 | 5mo ago |