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

MME-RealWorld

Benchmarks

Task NameDataset NameSOTA ResultTrend
Multimodal UnderstandingMME-RealWorld Lite
Overall Score67.3
34
PerceptionMME-RealWorld Lite
Overall Score59
29
Real-world Multimodal UnderstandingMME-RealWorld Lite
Lite Score54.9
25
Multimodal UnderstandingMME-RealWorld Chinese
Accuracy64.3
25
Multimodal UnderstandingMME-RealWorld English
Accuracy59.9
25
ReasoningMME-RealWorld Lite
OCR Score84
20
Real-world UnderstandingMME-RealWorld EN
Score64
20
Multimodal Question AnsweringMME-RealWorld-Lite 1.0 (test)
Perception (AD) Acc57.7
19
General Visual ReasoningMME-RealWorld-Lite
Accuracy73.06
17
Multimodal EvaluationMME-RealWorld
Accuracy71.2
15
Fine-grained visual reasoningMME Realworld Lite
Avg@155.8
12
ReasoningMME-RealWorld Lite (test)
OCR76
12
Remote Sensing Visual Question AnsweringMME-RealWorld-RS
Position Score58.15
11
General Perception and ReasoningMME-RealWorld Lite
Overall Accuracy54.3
11
Multimodal EvaluationMME-RealWorld Lite
Score57.8
10
Real-world Visual Question AnsweringMME-RealWorld-Lite (MMERW)
Accuracy44.6
8
Fine-grained PerceptionMME-RealWorld Lite
Score51.49
6
General Visual Question AnsweringMME-RealWorld en
Score63.2
6
PerceptionMME-RealWorld Lite (test)
OCR83.6
3
Multimodal EvaluationMME-RealWorld zero-shot
Zero-shot Accuracy48.03
2
Showing 20 of 20 rows