Mathematical Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
91.34Accuracy
5
Feb 26, 2026
95.5Accuracy
5
Feb 26, 2026
94.3Accuracy
5
Feb 26, 2026
57.5Score
5
Feb 26, 2026
2.14Speedup
5
Feb 26, 2026
71.9Accuracy
5
Feb 26, 2026
24.68Exact Match (EM)
4
Jul 3, 2026
15.9Query Prefill Latency (ms)
4
Jul 1, 2026
2.47Accuracy
4
Jul 1, 2026
81Accuracy (GSM8K)
4
Jul 1, 2026
0.32Forward Pass Time (s)
4
Jun 30, 2026
403.45TPS
4
Jun 30, 2026
834.52Tokens Per Second (TPS)
4
Jun 30, 2026
64.4Exact Match
4
Jun 25, 2026
27.5Accuracy
4
Jun 23, 2026
18.01Accuracy
4
Jun 23, 2026
49.8Accuracy
4
Jun 23, 2026
94.54Accuracy
4
Jun 19, 2026
435p@6=0
4
Jun 19, 2026
83p@6=0
4
Jun 19, 2026
5.16Latency (s)
4
Jun 16, 2026
76.7Accuracy
4
Jun 16, 2026
72.5Clean Accuracy
4
Jun 16, 2026
66.9Accuracy
4
Jun 15, 2026
295.2Inference Time
4
Jun 11, 2026