Commonsense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
70.9Accuracy
26
Jul 3, 2026
62.67HellaSwag Zero-shot Accuracy
25
Jul 7, 2026
82.81PIQA Accuracy
25
Jun 26, 2026
41.47ARC-C Accuracy
25
May 4, 2026
83.6ARC-e Accuracy
25
Jun 30, 2026
70.84Accuracy
25
Feb 26, 2026
76.13Accuracy (%)
24
Jun 1, 2026
80.71Accuracy
24
May 7, 2026
80.66WinoGrande Accuracy (WG)
24
Apr 21, 2026
79.8First-Token Accuracy
24
Apr 6, 2026
93.2First-Token Accuracy
24
Apr 6, 2026
91.4First-Token Accuracy
24
Apr 6, 2026
88ARC-e Accuracy
24
Mar 10, 2026
94.8OBQA
24
Feb 26, 2026
85.87Accuracy
23
Jun 26, 2026
79.1Accuracy
23
Feb 26, 2026
65.2Accuracy (5-6)
22
Jun 26, 2026
73.88Winogrande Score
22
May 21, 2026
75.91BoolQ Accuracy
22
Apr 9, 2026
57Accuracy
22
Jul 3, 2026
77.74Score
22
Jul 3, 2026
69.85Accuracy
21
May 6, 2026
29.8Accuracy
21
Apr 10, 2026
73.95Accuracy
21
Jun 26, 2026
95.6Accuracy
21
Feb 26, 2026