Commonsense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
99.21Accuracy
1,896
May 11, 2026
7,364Accuracy
1,581
Jul 9, 2026
87.4HellaSwag Accuracy
897
Jun 29, 2026
94.9Accuracy
757
May 18, 2026
85.3Accuracy
453
May 28, 2026
89.99Accuracy
400
Jul 9, 2026
96Accuracy
366
Feb 26, 2026
93Avg Score
321
May 8, 2026
88BoolQ Accuracy
283
Jul 1, 2026
97.35Accuracy
259
Jun 16, 2026
96.56Accuracy
249
Jul 8, 2026
89.69BoolQ Accuracy
245
Jul 2, 2026
96.3Accuracy
225
Jun 30, 2026
95.7Accuracy
208
May 14, 2026
91.2CSQA Accuracy
195
May 26, 2026
89.2Accuracy
187
May 19, 2026
89.85Accuracy
183
Jun 26, 2026
88.1Accuracy
164
Jun 17, 2026
86.55Accuracy
146
May 12, 2026
91.2Accuracy
136
Mar 4, 2026
89.4Accuracy
123
Jun 4, 2026
83.49Accuracy
119
May 29, 2026
86.6Accuracy (pass@1)
108
May 4, 2026
91.2Accuracy
108
May 12, 2026
86.63Accuracy
106
Jul 8, 2026