Commonsense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
0.725Accuracy
8
Feb 26, 2026
65.1Winogrande ET Accuracy
8
Feb 26, 2026
60.6Accuracy (Before Attack)
8
Feb 26, 2026
67.21Factuality (MAUVE)
8
Feb 26, 2026
0.968Accuracy
8
Feb 26, 2026
91.66Accuracy
8
Feb 26, 2026
91.28Accuracy
8
Feb 26, 2026
42.88Accuracy
8
Feb 26, 2026
83.14Accuracy
8
Feb 26, 2026
75.4Accuracy
8
Feb 26, 2026
79.2Accuracy
8
Feb 26, 2026
60.34Average Accuracy
7
Jun 9, 2026
32.8HellaSwag Accuracy
7
Jun 1, 2026
36.11ARC-Challenge Accuracy
7
May 19, 2026
54.4Normalized Accuracy
7
Apr 23, 2026
92.6ASRr
7
Apr 14, 2026
34OpenBookQA
7
Apr 3, 2026
42.6Overall Score
7
Mar 20, 2026
83.3Accuracy
7
Mar 18, 2026
90.2PIQA Accuracy
7
Feb 27, 2026
0.534ARC-c Accuracy
7
Feb 26, 2026
71StrategyQA Accuracy
7
Feb 27, 2026
74Accuracy
7
Feb 26, 2026
0.714AUC
7
Feb 26, 2026
68Accuracy (EN)
6
Jun 11, 2026