Mathematical Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
78Accuracy
3
Feb 26, 2026
84Accuracy
3
Feb 26, 2026
84.67Accuracy
3
Feb 26, 2026
75Accuracy
3
Feb 26, 2026
99.5ACC
3
Feb 26, 2026
84.46Accuracy
3
Feb 26, 2026
28.9GSM8K Accuracy
3
Feb 26, 2026
237.96Throughput (tokens/s)
3
Feb 26, 2026
78Answer Accuracy
3
Feb 26, 2026
94Answer Accuracy
3
Feb 26, 2026
40.16ACC (%)
3
Feb 26, 2026
89SAT Math Percentile
3
Feb 26, 2026
0.539Overall Accuracy
3
Feb 26, 2026
77.3Best Single Router Performance
2
Jul 7, 2026
0.96Best Single Router Performance
2
Jul 7, 2026
68.41GSM8K Accuracy (%)
2
Jul 1, 2026
32.2Accuracy (Full)
2
Jul 1, 2026
0.48Correlation Coefficient (s)
2
Jun 30, 2026
0.93s Score
2
Jun 30, 2026
100Optimization Steps
2
Jun 17, 2026
0.886Sel-AUC
2
Jun 17, 2026
94.12Pass@2
2
Jun 9, 2026
73.67Pass@2
2
Jun 9, 2026
89.6Pass@1 Accuracy
2
Jun 8, 2026
91Pass@1
2
Jun 8, 2026