Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Microscopic Reasoning on MicroVQA
Loading...
59.2
Accuracy
Gemini-2.5-Pro
37.984
43.492
49
54.508
May 21, 2026
Accuracy
Updated 12d ago
Evaluation Results
Method
Method
Links
Accuracy
Gemini-2.5-Pro
Model Type=Closed-Source
2026.05
59.2
Gemini-2.5-Flash
Model Type=Closed-Source
2026.05
58.6
GPT-5
Model Type=Closed-Source
2026.05
57
MAESTRO
Model Type=Ours
2026.05
53
Qwen3-VL-32B
Model Type=Open-Source...
2026.05
52.6
Kimi-K2.5
Model Type=Open-Source...
2026.05
51.6
GLM-4.6V
Model Type=Open-Source...
2026.05
51
PixelReasoner
Model Type=Think with...
2026.05
50.8
Visual-ARFT
Model Type=Think with...
2026.05
50.4
VTS-V
Model Type=Think with...
2026.05
49.4
VisionReasoner
Model Type=Think with...
2026.05
49
DeepEyes
Model Type=Think with...
2026.05
48.8
Thyme
Model Type=Think with...
2026.05
48.8
GPT-4o
Model Type=Closed-Source
2026.05
47.8
Chain-of-Focus
Model Type=Think with...
2026.05
47.4
MathCoder-VL
Model Type=Think with...
2026.05
44
VTOOL-R1
Model Type=Think with...
2026.05
43.8
DeepEyes-v2
Model Type=Think with...
2026.05
41.4
Direct Answering
Model Type=Open-Source...
2026.05
40.8
Untrained Model
Model Type=Open-Source...
2026.05
38.8
Feedback
Search any
task
Search any
task