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

VisualPuzzles

Benchmarks

Task NameDataset NameSOTA ResultTrend
Visual ReasoningVisualPuzzles
Overall Score75.25
12
Visual ReasoningVisualPuzzles OOD (test)
Overall Accuracy47.95
8
Multimodal Visual Logic ReasoningVisualPuzzles
Mean@558.2
8
Visual logicVisualPuzzles
Top-1 Accuracy43.15
7
Multimodal ReasoningVisualPuzzles (out-of-distribution)
Overall Accuracy33.9
6
Showing 5 of 5 rows