Commonsense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
76.09Accuracy
103
May 12, 2026
77.51Average Accuracy (Zero-shot)
99
Jul 9, 2026
72.8Average Accuracy
98
Jun 23, 2026
74.9Accuracy
94
Jun 24, 2026
78.95Accuracy
90
Jul 7, 2026
70.47CMQA Accuracy (%)
88
Apr 29, 2026
0.959HellaSwag Accuracy
87
Jun 9, 2026
73.88Accuracy
87
Feb 26, 2026
92.66Accuracy
85
May 25, 2026
88.22Accuracy (8 Sub-Tasks)
80
Jun 4, 2026
90.1Accuracy
73
Jul 3, 2026
72.04Avg Accuracy
72
Apr 3, 2026
95.3Accuracy
68
Jun 23, 2026
78.94HS Accuracy
66
Jun 2, 2026
90.92Normalized Accuracy
66
Jun 29, 2026
82.21Accuracy
64
Apr 21, 2026
86.86HellaSwag Score
62
May 11, 2026
85.7Accuracy
62
Jun 17, 2026
90Accuracy
62
Apr 9, 2026
82.35Accuracy
61
Jun 26, 2026
93Accuracy
61
Jun 1, 2026
68.6Zero-Shot Accuracy
60
Jun 11, 2026
85Accuracy
57
Jun 2, 2026
95.45HellaSwag Score
55
Mar 20, 2026
79.95Accuracy
55
May 13, 2026