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

CVBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Self-evaluationCVBench
AUROC0.747
36
VisionCVBench
CVBench Score81.5
23
Visual PerceptionCVBench 2D
Score79.09
22
Vision UnderstandingCVBench-2D
Accuracy77.76
22
Computer Vision BenchmarkingCVBench
Accuracy83
19
Spatial ReasoningCVBench
2D Relationship Score96.31
15
Visual PerceptionCVBench 3D
Score89.1
13
Visual ReasoningCVBench
Score76.9
12
Spatial ReasoningCVBench
Count Score70.6
12
Vision-Language ReasoningCVBench
Accuracy86.16
12
Multimodal UnderstandingCVBench
Accuracy64.25
11
PerceptionCVBench (test)
Accuracy87.6
11
Video UnderstandingCVBench 83 (test)
Average Score69.1
10
Spatial UnderstandingCVBench-3D
Overall Score83.1
9
Spatial UnderstandingCVBench-2D
Overall Score79.4
9
Multiple-choice Question AnsweringCVBench
Accuracy79.3
8
Image VQACVBench
Score75.5
7
3D TaskCVBench
Accuracy84.52
7
Visual Question AnsweringCVBench 3D
Accuracy87
7
Visual Question AnsweringCVBench 2D
Accuracy77.7
7
General Image-Text UnderstandingCVBench
Accuracy86.32
6
Visual PerceptionCVBench
2D Score76.6
5
Spatial reasoningCVBench
Score57.3
4
Complex Reasoningcvbench
Accuracy54.9
4
2D Vision ReasoningCVBench 2D
Accuracy58.5
4
Showing 25 of 27 rows