Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Compositional Reasoning on MMStar
Loading...
64.7
Accuracy
GPT-4o
45.044
50.147
55.25
60.353
Apr 12, 2026
Accuracy
Updated 5d ago
Evaluation Results
Method
Method
Links
Accuracy
GPT-4o
Paradigm=Zero-Shot VLM...
2026.04
64.7
VL-Rethinker
Paradigm=Tool-use & RL...
2026.04
63.2
Vision-R1
Paradigm=Tool-use & RL...
2026.04
62.67
SCF-VR
Paradigm=Latent Reason...
2026.04
60.82
ICoT
Paradigm=Explicit CoT...
2026.04
60.4
Monet
Paradigm=Latent Reason...
2026.04
60.33
DMLR
Paradigm=Latent Reason...
2026.04
60.27
Laser
Paradigm=Latent Reason...
2026.04
60.1
Qwen2.5-VL-7B
Paradigm=Zero-Shot VLM...
2026.04
59.7
LLaVA-OneVision
Paradigm=Zero-Shot VLM...
2026.04
59.13
DeepEyes
Paradigm=Tool-use & RL...
2026.04
58.73
CCoT
Paradigm=Explicit CoT...
2026.04
58.7
LVR
Paradigm=Latent Reason...
2026.04
57.93
Multimodal CoT
Paradigm=Explicit CoT...
2026.04
57.9
InternVL3.5-8B
Paradigm=Zero-Shot VLM...
2026.04
53.33
PAPO
Paradigm=Tool-use & RL...
2026.04
45.8
Feedback
Search any
task
Search any
task