Math Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
61.4Pass@1
3
Feb 26, 2026
83.2Pass@1 Accuracy
3
Feb 26, 2026
3.63Speedup
2
Jul 1, 2026
57.68NumGLUE Accuracy
2
Jun 9, 2026
37.2BA @20
2
May 29, 2026
95.2BA Accuracy @20
2
May 29, 2026
56.5BA @20 Accuracy
2
May 29, 2026
4.79Pass@1
2
May 26, 2026
0.63Pass@1
2
May 26, 2026
9.27Pass@1
2
May 26, 2026
3.95Tau (Mean Accepted Tokens)
2
Apr 8, 2026
3.65Tau (Accepted Tokens)
2
Apr 8, 2026
3.81Tau
2
Apr 8, 2026
3.82Mean Accepted Tokens
2
Apr 8, 2026
23.3Pass@1 Rate
2
Feb 26, 2026
55.65Accuracy
2
Feb 26, 2026