Mathematical Reasoning on AIME25, HMMT-Feb, HMMT-Nov, MATH500, Minerva (test)
61.2AIME25 AccuracyTeacher
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Teacher2026.05 | 61.2 | 42.4 | 57.9 | 98 | 51.7 | 62.2 | |
| Fixed-prefix OPDPrefix Length=40962026.05 | 31.6 | 19.3 | 20.4 | 86.8 | 38.8 | 39.4 | |
| Ours (trajectory-specific release rule)Release Rule=trajectory-specific2026.05 | 31.5 | 19.1 | 21.9 | 89.4 | 38.8 | 40.1 | |
| ExOPD2026.05 | 30.8 | 16.4 | 17.4 | 89.2 | 41.2 | 39 | |
| Fixed-prefix OPDPrefix Length=81922026.05 | 30.7 | 17.4 | 18.4 | 88.8 | 37.3 | 38.6 | |
| Fixed-prefix OPDPrefix Length=20482026.05 | 30.6 | 18.5 | 20.9 | 89.2 | 37.3 | 39.3 | |
| Fixed-prefix OPDPrefix Length=10242026.05 | 30.1 | 18.9 | 20.6 | 87.8 | 35.9 | 38.7 | |
| Random release2026.05 | 29.2 | 18 | 18.9 | 88 | 36 | 38 | |
| OPD2026.05 | 28.4 | 16.2 | 15.9 | 86.2 | 37.1 | 36.8 | |
| Student2026.05 | 11.4 | 6.6 | 4.5 | 74 | 27.9 | 24.9 |