Mathematical Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
6.02Speedup
16
Jun 11, 2026
97.1Accuracy
16
Jun 16, 2026
39.7AIME25 Accuracy
16
Jun 2, 2026
86.6Accuracy on MATH500
16
May 29, 2026
4.92Cumulative Rollouts (M)
16
May 27, 2026
90.33Accuracy (%)
16
May 26, 2026
0.964Accuracy
16
May 26, 2026
78Knowledge Recall
16
May 26, 2026
97.6LLM Trust Score
16
May 22, 2026
95.3LLM Trust Score
16
May 22, 2026
98.5LLM Trust Score
16
May 22, 2026
99.91LLM Trust Score
16
May 22, 2026
99.97LLM Trust Score
16
May 22, 2026
99.97LLM Trust Score
16
May 22, 2026
35.29Reference Score/Rate
16
May 22, 2026
33.4S_ref
16
May 22, 2026
58.1Initial Task Score
16
May 22, 2026
61.79Throughput
16
May 22, 2026
76.6Average Pass@1
16
May 19, 2026
80.7Pass@1
16
May 19, 2026
80.2Mean Harmonic Score (MH)
16
May 13, 2026
87.9GSM Accuracy
16
May 13, 2026
95.8M500 Score
16
May 13, 2026
62.3A25 Score
16
May 13, 2026
53.3A24 Score
16
May 13, 2026