Mathematical Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
37.1Accuracy
4
Jun 9, 2026
38.3Accuracy
4
Jun 9, 2026
70.7Accuracy
4
Jun 9, 2026
84Average Score @16
4
Jun 9, 2026
35Avg@16
4
Jun 9, 2026
74.5Avg@4 Score
4
Jun 9, 2026
52.9Accuracy (avg@4)
4
Jun 9, 2026
95.8Average @ 4
4
Jun 9, 2026
91.9Avg@32
4
Jun 9, 2026
99.9Avg@30 Accuracy
4
Jun 8, 2026
82.2Avg Score (AIME 2025)
4
Jun 8, 2026
85.3Avg@30
4
Jun 8, 2026
30.62Average Latency (MATH)
4
Jun 4, 2026
10.67Average Latency
4
Jun 4, 2026
76.82S-Format Score
4
Jun 4, 2026
72.63S-Format Compliance
4
Jun 4, 2026
86.3Pass@1 Success Rate (Last)
4
Jun 4, 2026
32.7Pass@1 (Last Sample)
4
Jun 4, 2026
86.5Pass@1 Accuracy
4
Jun 4, 2026
50Accuracy (Pass@1)
4
Jun 4, 2026
72.3Pass@1 Accuracy
4
Jun 4, 2026
64Pass@1 Accuracy
4
Jun 4, 2026
33.92TPS
4
Jun 4, 2026
29.75Throughput (TPS)
4
Jun 4, 2026
32.44Throughput (TPS)
4
Jun 4, 2026