Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Perception-intensive Reasoning on MME-RealWorld-Lite
Loading...
55.13
Score
Starve to Perceive
26.2492
33.7471
41.245
48.7429
May 18, 2026
Score
Updated 15d ago
Evaluation Results
Method
Method
Links
Score
Starve to Perceive
Visual Bandwidth Const...
2026.05
55.13
Qwen2.5-VL + BA-SFT + VanillaRL
Visual Bandwidth Const...
2026.05
53.78
DeepEyes
Visual Bandwidth Const...
2026.05
50.81
ChainOfFocus
Visual Bandwidth Const...
2026.05
48.98
Starve to Perceive
Visual Bandwidth Const...
2026.05
47.37
PixelReasoner
Visual Bandwidth Const...
2026.05
47.21
Qwen2.5-VL + BA-SFT
Visual Bandwidth Const...
2026.05
46.95
Qwen2.5-VL + BA-SFT + VanillaRL
Visual Bandwidth Const...
2026.05
46.69
GPT-4o
Visual Bandwidth Const...
2026.05
46.4
LLaVA-OneVision
Visual Bandwidth Const...
2026.05
43.7
Qwen2.5-VL
Visual Bandwidth Const...
2026.05
43
DeepEyes
Visual Bandwidth Const...
2026.05
40.85
Qwen2.5-VL + BA-SFT
Visual Bandwidth Const...
2026.05
39.55
ChainOfFocus
Visual Bandwidth Const...
2026.05
37.83
Mini-o3
Visual Bandwidth Const...
2026.05
32.15
PixelReasoner
Visual Bandwidth Const...
2026.05
29.6
Qwen2.5-VL
Visual Bandwidth Const...
2026.05
27.41
Mini-o3
Visual Bandwidth Const...
2026.05
27.36
Feedback
Search any
task
Search any
task