Fine-grained Visual Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
90Overall Score
20
Jul 9, 2026
87.9Overall Score
20
Jul 9, 2026
57Overall Score
20
Jul 9, 2026
88.5FSP Score
19
Jul 9, 2026
95.7Overall Accuracy
14
Jul 2, 2026
89.5Avg@8 Overall
13
Apr 10, 2026
85.9Average Score
12
Jul 2, 2026
94.5FSP Score
12
Jul 2, 2026
55.8Avg@1
12
Apr 10, 2026
70.2Score
10
Jun 25, 2026
84.5Score
10
Jun 25, 2026
87.8Score
10
Jun 25, 2026
91.6Score
10
Jun 25, 2026
89Accuracy
8
Feb 26, 2026
77.5Score
6
Jun 25, 2026
57.8Score
6
Jun 25, 2026
45.2AR
4
May 26, 2026