Arithmetic Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
12.4Δ Accuracy
6
Feb 26, 2026
99.64Source Library Accuracy
5
Jun 19, 2026
87.79Accuracy (Exact Match)
5
May 11, 2026
52.06Accuracy
5
Mar 26, 2026
57.79Accuracy
5
Mar 26, 2026
79.65Accuracy
5
Mar 26, 2026
46.6Pass@1
5
Mar 4, 2026
71Accuracy
4
Apr 9, 2026
41.3Success Rate
4
Mar 4, 2026
94.4Success Rate
4
Mar 4, 2026
89.9Accuracy
4
Feb 26, 2026
72.7Accuracy
4
Feb 26, 2026
77.5Accuracy
4
Feb 26, 2026
100Accuracy
3
Jul 3, 2026
90Accuracy
3
Jul 3, 2026
92Accuracy
3
Jul 3, 2026
88Accuracy
3
Jul 3, 2026
100Accuracy
3
Jun 4, 2026
98.81Accuracy
3
Jun 4, 2026
87.84Accuracy
3
Jun 4, 2026
593Correct Count
3
Jun 4, 2026
1,885Correct Count
3
Jun 4, 2026
73.9Accuracy
3
Feb 27, 2026
62.1Accuracy (format-specific)
2
Feb 27, 2026
33.5Accuracy (format-specific prompt)
2
Feb 27, 2026