Multi-image Understanding
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
55.8Accuracy
65
Apr 14, 2026
74.5Score
41
May 1, 2026
81.7Accuracy
30
Mar 10, 2026
71.8Score
23
Feb 27, 2026
68Score
23
Feb 27, 2026
72.42Accuracy
22
Mar 10, 2026
57.3Temporal Multi-Image Score (Task T)
21
Mar 6, 2026
68Accuracy
21
Feb 26, 2026
72.1Score
19
Feb 26, 2026
86.1Score
19
Feb 26, 2026
62.3Accuracy
17
Feb 26, 2026
60.8Accuracy
12
Feb 26, 2026
56.8Accuracy
12
Feb 26, 2026
55.6Accuracy
11
Feb 26, 2026
67.4Accuracy
11
Feb 26, 2026
52.6Accuracy
11
Feb 26, 2026
87.3Accuracy
9
Feb 26, 2026
54.2Accuracy
7
Feb 26, 2026
36.9Accuracy
6
Feb 26, 2026
79.7Score
5
May 1, 2026
61.18Accuracy
5
Apr 14, 2026
66.36Accuracy
2
Jun 23, 2026