Common Sense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
67.81Accuracy
15
Feb 26, 2026
56.24Average Accuracy
14
Jun 2, 2026
86.73Pass@1
14
Feb 26, 2026
47.3ARC Accuracy
13
Apr 28, 2026
39.3ARC Score
12
Apr 28, 2026
88.43Accuracy
12
Apr 7, 2026
83.61Accuracy
12
Apr 7, 2026
78.09Accuracy (ACC)
12
Apr 7, 2026
74.9Accuracy
12
Jun 16, 2026
66.87Character Accuracy
11
Mar 18, 2026
47.9ARC Accuracy
10
Apr 28, 2026
62Accuracy
10
Mar 24, 2026
88.24Accuracy
9
Jul 3, 2026
83.99Accuracy
9
Jul 3, 2026
82.55Accuracy
9
Jul 3, 2026
78.82Accuracy
9
Jul 3, 2026
75.33Average Accuracy
9
Feb 26, 2026
25.89PPL (WikiText)
8
May 26, 2026
34ARC Challenge Accuracy
8
Feb 26, 2026
61.5Accuracy
8
Feb 26, 2026
18.22Hit Rate @1
7
Feb 26, 2026
19.02H@1
7
Feb 26, 2026
85.99Accuracy
6
Mar 30, 2026
79.65Accuracy
6
Mar 30, 2026
42.91Average Score (8 tasks)
6
Mar 25, 2026