Mathematical Reasoning on In-Domain Reasoning Suite (MATH, Olympiad, AMC, AIME)
94.4MATH Score
GSPO + LIE
Updated 1mo ago
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
2026.02 | 94.4 | 67 | 85.3 | 57.7 | 46.4 | 70.2 | |
2026.02 | 94 | 68.1 | 82 | 54.2 | 42.5 | 68.2 | |
2026.02 | 82.8 | 51.9 | 60.4 | 24.2 | 19.4 | 47.7 | |
2026.02 | 60.8 | 28.1 | 30.3 | 4.5 | 4.4 | 25.6 | |
2026.02 | 55.8 | 23.1 | 28.2 | 3.8 | 2.3 | 22.6 | |
2026.02 | 23.4 | 9 | 10.4 | 1.1 | 0.6 | 8.9 |