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

Geo3K

Benchmarks

Task NameDataset NameSOTA ResultTrend
Geometry problem solvingGeo3K
Top-1 Accuracy98.2
41
Multimodal ReasoningGeo3K
Accuracy45.59
21
Mathematical ReasoningGeo3k
Accuracy52
18
Multimodal ReasoningGeo3k
Pass@162.23
17
Geometric ReasoningGeo3K
Accuracy55.2
16
Mathematical & Geometric ReasoningGeo3k
Accuracy@856.6
16
Multimodal Geometry ReasoningGeo3K
Accuracy49.75
11
Visual ReasoningGeo3K
Accuracy56.74
10
Visual Math ReasoningGeo3K
Accuracy51.58
9
Multimodal ReasoningGEO3K (test)
Accuracy70.1
7
Multimodal Mathematical ReasoningGeo3K
Accuracy50.58
6
Showing 11 of 11 rows