Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

MathVision

Benchmarks

Task NameDataset NameSOTA ResultTrend
Visual Mathematical ReasoningMathVision
Accuracy92.7
186
Multimodal Math ReasoningMathVision
Accuracy86
183
Mathematical ReasoningMathVision
Accuracy75.95
144
Multimodal reasoningMathVision
Accuracy57.9
102
Mathematical reasoningMathVision (test)
Accuracy71.9
53
Multimodal mathematical reasoningMathVision (test)
Accuracy60.3
47
Multi-modal ReasoningMathVision (test)
Accuracy (%)47.7
45
Mathematical Visual Question AnsweringMathVision
Accuracy73.3
34
Mathematical ReasoningMathVision Mini
Score47.36
25
STEM ReasoningMathVision
Accuracy69.96
23
Mathematical ReasoningMathVision
AUC (%)95.68
21
Mathematical ReasoningMathVision MVisionm
Accuracy41.3
18
Step-wise VerificationMathVision
Macro F161.7
18
Multimodal Mathematical ReasoningMathVision
Mean@579.5
16
Visual Mathematical ReasoningMathVision (test)
Score51.8
16
Mathematics ReasoningMathVision mini
Accuracy60.54
15
Mathematical ReasoningMathVision (testmini)
Accuracy29.9
13
Visual Mathematical ReasoningMathVision
BoN@8 Accuracy43.6
12
Mathematical ReasoningMathVision
Score63.5
11
Multimodal mathematical reasoningMathVision
Pass@1 Accuracy58.75
11
CoT Length PredictionMathVision m
rMAE0.2934
10
Fuel level estimationMathVision m
rMAE11.86
10
General Perception and ReasoningMathVision
Score58.6
10
Mathematical ReasoningMathVision
Top-1 Accuracy72
9
Visual Math ReasoningMathVision
Score72
9
Showing 25 of 37 rows