Commonsense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
79Accuracy
6
Feb 26, 2026
91.7Accuracy
5
Jun 2, 2026
29.6OBQA Score
5
May 22, 2026
53.33ARC-Challenge Accuracy
5
May 19, 2026
56Accuracy
5
Apr 24, 2026
52.3HellaSwag
5
Apr 16, 2026
73.6HellaSwag
5
Apr 15, 2026
7.02Perplexity
5
Mar 30, 2026
81.1Winogrande Accuracy
5
Feb 27, 2026
73.88Accuracy
5
Feb 26, 2026
54.77Accuracy
5
Feb 26, 2026
20Accuracy
5
Feb 26, 2026
40.4Correctness Rate
5
Feb 26, 2026
66.17AUC (%)
5
Feb 26, 2026
79.82Accuracy
5
Feb 26, 2026
92.5aNLI
5
Feb 26, 2026
95.6Accuracy
5
Feb 26, 2026
0.94AUC
5
Feb 26, 2026
78.8Accuracy (XS)
5
Feb 26, 2026
34.7Accuracy
4
Jul 3, 2026
78.95Accuracy
4
Jun 30, 2026
45.4Best Accuracy
4
Jun 23, 2026
26Normalized Accuracy
4
Jun 16, 2026
51.67Accuracy (Clean)
4
Jun 16, 2026
59.9Accuracy
4
Jun 11, 2026