| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| OlympiadBench | TDA-RC | Accuracy12.8 | 18 | 11d ago | |
| Math500 512 tokens | d-TreeRPO | Pass@1 Accuracy46.3 | 15 | 1mo ago | |
| Math500 256 tokens | d-TreeRPO | Pass@1 Accuracy41.2 | 15 | 1mo ago | |
| MATH 500 | NPG-Muse-8B | Accuracy85.5 | 6 | 1mo ago | |
| AIME 25 | NPG-Muse-8B | AIME 25 Accuracy19.9 | 6 | 1mo ago | |
| AIME All Non-Easy | QA+ | Recovery Rate22.2 | 3 | 13d ago | |
| AIME Med. + Hard | Bit-Limited Chain-of-Thought (BL-CoT) | Recovery Rate10.8 | 3 | 13d ago |