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

MMAR

Benchmarks

Task NameDataset NameSOTA ResultTrend
Audio ReasoningMMAR (test)
Average Score74.4
57
Audio Question AnsweringMMAR
Average Score80.4
47
Audio ReasoningMMAR
Average Accuracy83.7
38
Audio-Language ReasoningMMAR 1.0 (test)
Accuracy70.3
27
Audio UnderstandingMMAR (comprehensive evaluation)
Sound Score62.4
25
Multimodal Audio ReasoningMMAR
Mean Score63.5
22
Audio UnderstandingMMAR (test)
Performance67.1
20
Audio UnderstandingMMAR
Average Score83.7
15
Audio Perception and ReasoningMMAR within CAFE framework (overall)
Perception Accuracy63.51
13
Audio Understanding / Audio ReasoningMMAR
Accuracy61.4
13
Massive Multimodal Audio ReasoningMMAR
Normalized Judge Score56.4
9
Audio ReasoningMMAR Agent Track
Accuracy77.4
8
Audio UnderstandingMMAR
Speech Score45.24
5
Audio ReasoningMMAR N=1,000
Accuracy53.6
5
Audio Understanding & ReasoningMMAR
Score71.9
3
Dense Audio CaptioningMMAR
MMAR Score46.4
3
Audio QAMMAR (test)
Accuracy68.1
2
Showing 17 of 17 rows