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

MME-RW

Benchmarks

Task NameDataset NameSOTA ResultTrend
Fine-grained Visual ReasoningMME-RW-Lite ID
Overall Score57
20
Real-world Multimodal EvaluationMME-RW (test)
Overall Score62.9
15
Multimodal UnderstandingMME-RW-en (test)
Overall Score71.4
15
Visual ReasoningMME-RW Chinese
Accuracy77.7
14
Showing 4 of 4 rows