Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Commonsense Reasoning on CSQA (Accuracy)

91.2CSQA Accuracy

Previous SoTA

65.61672.25878.985.542Mar 21, 2022Nov 22, 2022Jul 27, 2023Mar 29, 2024Dec 1, 2024Aug 4, 2025Apr 8, 2026
Updated 9d ago

Evaluation Results

MethodLinks
2022.03
91.2
2026.04
87.63
2026.04
87.47
2026.04
87.47
2026.04
87.47
2026.04
87.24
2026.04
85.88
2026.04
85.6
2026.04
84.95
2026.04
84.9
2026.04
84.9
2026.04
83.9
2026.04
83.4
2026.04
83.4
2026.02
82.8
2026.04
82.6
2026.02
82.56
2026.02
82.47
2026.04
82.3
2026.04
82.2
2026.04
81.9
2022.03
81.5
2026.04
81.5
2026.02
81.45
2026.04
81.2
2026.04
81.1
2026.04
81
2026.04
81
2026.02
80.85
2022.03
80.7
2026.02
79.85
2022.10
79.5
2026.04
79.4
2026.04
79.1
2026.02
79.05
2022.03
79
2022.03
79
2026.02
78.78
2026.04
78.7
2026.02
78.45
2026.03
78.3
2026.03
78.2
2026.03
78.1
2026.02
77.9
2026.03
77.8
2026.03
77.8
2026.03
77.8
2026.03
77.7
2026.03
77.6
2026.03
77.6
2026.03
77.6
2026.02
77.42
2026.02
77.42
2026.02
77.36
2026.02
76.82
2026.03
76.7
2026.02
76.4
2026.03
76.4
2026.02
75.18
2026.03
75
2026.03
75
2026.02
74.53
2022.10
74.4
2026.02
74.33
2026.02
74.12
2022.10
73.5
2026.02
73.12
2026.02
73.01
2026.02
73.01
2022.10
72.6
2026.02
72.48
2026.03
72.4
2026.03
71.8
2026.03
71.4
2026.02
71.33
2026.03
71.3
2026.03
71.2
2026.03
71.2
2026.03
71
2026.03
70.8
2026.03
70.7
2026.03
70.6
2026.03
70.6
2026.03
70.4
2026.02
70.22
2026.03
70.2
2026.02
70.15
2026.03
69.2
2026.03
69.2
2026.03
68.9
2026.03
68.7
2026.03
68.7
2026.03
68.2
2026.03
68.2
2026.02
68.11
2026.03
68.1
2026.03
67.9
2026.03
67.3
2026.03
66.9
2026.03
66.6
Showing 100 of 126 rows