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

Video Spatial Reasoning on SpaceVista-Bench

81.3Accuracy

Human

10.89229.17147.4565.729Oct 10, 2025
Updated 7d ago

Evaluation Results

MethodLinks
2025.10
81.3
2025.10
39.2
2025.10
35.4
33.8
2025.10
33.7
2025.10
31.1
2025.10
30.7
2025.10
29.5
2025.10
28.9
2025.10
28.8
2025.10
24.5
2025.10
24.2
2025.10
23.7
2025.10
23.6
2025.10
21.2
2025.10
13.6