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

VideoQA on LongVideoBench (Medium/Long/All Scores)

72.6Score (All Lengths)

GPT-5

49.61655.58361.5567.517Mar 31, 2026
Updated 18d ago

Evaluation Results

MethodLinks
2026.03
72.6--
2026.03
66.769.160.9
2026.03
6465.358.6
2026.03
63.9--
63.665.557.3
2026.03
6060.752.1
2026.03
56.3--
2026.03
56--
2026.03
52.1--
2026.03
50.54945.2