Long-context video understanding
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
66.7Accuracy
28
Feb 26, 2026
81.3Accuracy (w subs)
13
Feb 26, 2026
70.3Score
7
Mar 19, 2026
62.8Score
7
Mar 19, 2026
74Score
7
Mar 19, 2026
69.3Score
7
Mar 19, 2026
70Score
4
Jun 9, 2026
66Score
4
Jun 9, 2026
60.9Score
4
Mar 19, 2026