Mathematical Reasoning on MATH500, AIME24, and AIME25 Average (test)
2.5Delta Accuracy (pp)OPSDC
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OPSDCThinking-enabled=true, Math-reasoning=true, Model Scale=7-8B2026.05 | 2.5 | 22.8 | |
| Correct-onlyThinking-enabled=true, Math-reasoning=true, Model Scale=7-8B2026.05 | 1 | 29.2 |