Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Long Video Understanding on MLVU MCQ (test)

81.5Accuracy

Gemini 2.5 Pro

47.07656.01364.9573.887Mar 30, 2026
Updated 18d ago

Evaluation Results

MethodLinks
2026.03
81.5
2026.03
77.7
2026.03
75.7
2026.03
75.1
2026.03
69.1
2026.03
64
2026.03
63
2026.03
62.4
2026.03
60.6
2026.03
60.4
2026.03
60.2
2026.03
58.4
2026.03
57.6
2026.03
56.6
2026.03
56
2026.03
55.2
2026.03
53.8
2026.03
53.2
2026.03
52.8
2026.03
52.6
2026.03
52
2026.03
48.4