Logical reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
71Accuracy
5
May 29, 2026
58.2Accuracy
5
May 20, 2026
88.9Accuracy
5
May 7, 2026
99.5Accuracy
5
May 7, 2026
99.5Accuracy (RP Npred ≤ 30, 6 < δ ≤ 12 train)
5
May 7, 2026
130Puzzle Accuracy
5
May 4, 2026
92.69Average Score
5
Apr 30, 2026
92.21Accuracy
5
Mar 26, 2026
91.18Accuracy
5
Mar 26, 2026
99.97Accuracy
5
Mar 26, 2026
60.31Accuracy
5
Mar 26, 2026
72.78Accuracy
5
Mar 26, 2026
99.83Accuracy
5
Mar 26, 2026
100Accuracy
5
Mar 4, 2026
74.7Accuracy Pass@1
5
Mar 4, 2026
0.11Accuracy
5
Feb 26, 2026
85.2Accuracy
5
Feb 26, 2026
81.8Accuracy
5
Feb 26, 2026
37.3Avg H@3
5
Feb 26, 2026
59.6Avg H@3
5
Feb 26, 2026
29.6Avg Hits@3
5
Feb 26, 2026
78.33Accuracy
5
Feb 26, 2026
69.12Accuracy
5
Feb 26, 2026
98Final Accuracy
4
Jun 24, 2026
44.1Accuracy
4
Jun 23, 2026