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

Video Question Answering on RAVEN-QA Hard Queries

96.8Accuracy

VLM Only

-1.37624.11249.675.088Jun 23, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
96.8
2026.06
95.1
2026.06
92.7
2026.06
85.4
2026.06
85.4
2026.06
82.9
2026.06
82.9
2026.06
82.9
2026.06
82.9
2026.06
82.9
2026.06
82.9
2026.06
80.5
2026.06
80.5
2026.06
80.5
2026.06
80.5
2026.06
78
2026.06
78
2026.06
73.2
2026.06
73.2
2026.06
73.2
2026.06
73.2
2026.06
68.3
2026.06
68.3
2026.06
63.4
2026.06
63.4
2026.06
63.4
2026.06
63.4
2026.06
61
2026.06
61
2026.06
58.5
2026.06
58.5
2026.06
58.5
2026.06
56.1
2026.06
56.1
2026.06
56.1
2026.06
56.1
2026.06
53.7
2026.06
51.2
2026.06
51.2
2026.06
48.8
2026.06
48.8
2026.06
46.3
2026.06
46.3
2026.06
43.9
2026.06
43.9
2026.06
41.5
2026.06
41.5
2026.06
39
2026.06
36.6
2026.06
34.1
2026.06
34.1
2026.06
26.8
2026.06
24.4
2026.06
24.4
2026.06
22
2026.06
19.5
2026.06
17.1
2026.06
17.1
2026.06
14.6
2026.06
7.3
2026.06
4.9
2026.06
2.4
2026.06
2.4
2026.06
2.4