Multi-image Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
77.2Accuracy
112
Jun 30, 2026
63.57Accuracy
86
Jun 23, 2026
72.81Accuracy
39
Feb 26, 2026
81.71Accuracy
38
Apr 21, 2026
78.3Overall Score
28
Feb 27, 2026
68Accuracy
24
Feb 26, 2026
52.6Accuracy
21
Feb 26, 2026
79.3Accuracy
21
Feb 26, 2026
38.94Accuracy
21
Feb 26, 2026
8.96Single Scene Char Score
20
Feb 26, 2026
50.6Narration Score
12
May 28, 2026
44.7T-1 Score
9
Jul 1, 2026