| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| GPQA Diamond | Gemma 4 31B | Pass@1 Score84.3 | 20 | 19d ago | |
| Humanity's Last Exam 2,158 text-only | Avg@3 Score54.2 | 15 | 4mo ago | ||
| HLE (Humanity's Last Exam) text-only subset (val) | ReThinker | Inference Accuracy52.2 | 13 | 5mo ago | |
| XBench-DeepSearch 1.0 (test) | ReThinker | Inference Accuracy0.9 | 12 | 5mo ago | |
| GAIA text-only (val) | ReThinker | Inference Accuracy81.6 | 12 | 5mo ago | |
| HLE | Gemma 4 31B | Accuracy (HLE)26.5 | 5 | 19d ago |
