Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Computer Vision Reasoning on CV-Bench 3D
Loading...
82.9
Accuracy
Qwen2.5-VL 7B
59.396
65.498
71.6
77.702
Mar 4, 2025
May 17, 2025
Jul 31, 2025
Oct 13, 2025
Dec 27, 2025
Mar 11, 2026
May 25, 2026
Accuracy
Updated 1d ago
Evaluation Results
Method
Method
Links
Accuracy
Qwen2.5-VL 7B
Type=Generalist Baseline
2026.05
82.9
SpatialReasoner
Type=Specialist Reference
2026.05
80.3
w/ Entropy-Regularization
Evaluation Protocol=RL...
2026.05
78.8
Mini-o3
Mode=Zero-shot, Type=V...
2026.05
77.6
DeepEyes
Mode=Zero-shot, Type=V...
2026.05
76.7
vanilla RFT (Base)
Evaluation Protocol=RL...
2026.05
76.7
BLIP-3-4B
Model Scale=4B, Model...
2025.03
75.4
w/ Tool-Encourage Reward
Evaluation Protocol=RL...
2026.05
74.5
AKI-4B
Model Scale=4B, Model...
2025.03
71.8
Phi-3-Vision-4B
Model Scale=4B, Model...
2025.03
68.2
VILA-1.5-3B
Model Scale=3B, Model...
2025.03
60.3
Feedback
Search any
task
Search any
task