Visual Commonsense Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
82.57Accuracy
63
Feb 26, 2026
94Accuracy
54
Feb 26, 2026
95.7EM
27
Feb 27, 2026
57.67Accuracy
13
Feb 26, 2026
72.01Accuracy
10
Feb 26, 2026
67.1F1 Score
7
Mar 4, 2026
40.3Acc@50
7
Feb 26, 2026
59Q->A Accuracy
7
Feb 26, 2026
94.7Accuracy
7
Feb 26, 2026
80.6Q->A Accuracy
7
Feb 26, 2026
26.4Meteor Score
6
Feb 26, 2026
91.1CIDEr
5
Feb 26, 2026
36.93Accuracy
4
Feb 26, 2026
60.01Accuracy
4
Feb 26, 2026
79.8Q->A Accuracy
4
Feb 26, 2026
70.8Q -> A Accuracy
4
Feb 26, 2026