Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Video Reasoning on VBVR-Bench
Loading...
97.4
Overall Accuracy
Human
24.496
43.423
62.35
81.277
Mar 17, 2026
Mar 25, 2026
Apr 3, 2026
Apr 12, 2026
Apr 21, 2026
Apr 30, 2026
May 9, 2026
Overall Accuracy
Updated 22d ago
Evaluation Results
Method
Method
Links
Overall Accuracy
Human
Type=Baseline
2026.03
97.4
VBVR-Wan2.2 + CollabVR
VGM Cost (s)=10.91
2026.05
75.7
VBVR-Wan2.2 + Training-Free Ensemble
Type=Video Reasoning M...
2026.03
71.6
VBVR-Wan2.2 + Pass@4
VGM Cost (s)=14.80
2026.05
70.7
VBVR-Wan2.2 + Pass@2
VGM Cost (s)=7.40
2026.05
69.4
VBVR-Wan2.2
Type=Video Reasoning M...
2026.03
68.5
VBVR-Wan2.2
VGM Cost (s)=3.70
2026.05
67.1
VBVR-Wan2.2 + VideoTPO
VGM Cost (s)=11.10
2026.05
65
Sora 2
Type=Proprietary Video...
2026.03
54.6
Veo 3.1
Type=Proprietary Video...
2026.03
48
Runway Gen-4 Turbo
Type=Proprietary Video...
2026.03
40.3
Cosmos-Predict2.5 + CollabVR
VGM Cost (s)=10.91
2026.05
40.3
Wan2.2-I2V-A14B
Type=Open-source Video...
2026.03
37.1
Kling 2.6
Type=Proprietary Video...
2026.03
36.9
LTX-2
Type=Open-source Video...
2026.03
31.3
Cosmos-Predict2.5
VGM Cost (s)=3.70
2026.05
30.8
CogVideoX1.5-5B-I2V
Type=Open-source Video...
2026.03
27.3
HunyuanVideo-I2V
Type=Open-source Video...
2026.03
27.3
Feedback
Search any
task
Search any
task