Mathematical Reasoning on AIME 1983-2025 (combined)
13.28AccuracySD-E²
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SD-E²2026.01 | 13.28 | 16.7 | 2.6 | 710 | |
| GRPO-CFEE (count)reward_type=count2026.01 | 9.87 | 10.95 | 2.7 | 713 | |
| GRPO-CFLstrategy_type=single-trace2026.01 | 8.34 | — | — | 610 | |
| Multi-strategy promptstrategy_type=multi-trace2026.01 | 6.74 | 9.33 | 3.16 | 819 | |
| Single-strategy promptstrategy_type=single-trace2026.01 | 5.7 | — | — | 502 |