Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Audio Classification research benchmarks – Page 7 · SOTA2 Research

Audio Classification

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

InternVideo2_s2

98.6Top-1 Accuracy

3

Feb 26, 2026

Closed-ended audio classification suite (ESC50, FSD50K, DCASE, TUT, Beijing Opera, Vocal Sound) (test)

0.919ESC50 Accuracy

3

Feb 26, 2026

AudioSet AS-A Audio-only (test)

0.3mAP

3

Feb 26, 2026

DABS Speech 1.0 (downstream)

41.8Accuracy

3

Feb 26, 2026

AudioMNIST 40-150-10 (test)

95.23Reference Accuracy

2

Jun 23, 2026

NP (fine-tuning)

93.44Accuracy

2

Jun 9, 2026

AF-Next-Instruct

66.7Source Accuracy

2

Apr 14, 2026

44.4Accuracy

2

Mar 26, 2026

51.24Accuracy

2

Feb 26, 2026

ImageNet ESC-27 (val)

Audio-Image Linear

35Accuracy (1-shot)

2

Feb 26, 2026

ImageNet ESC-19 (val)

Audio-Image Linear

35.7Accuracy (1-shot)

2

Feb 26, 2026

Medley-solos-DB (test)

92.7Accuracy

2

Feb 26, 2026

78.7Accuracy

2

Feb 26, 2026

66.7F1 Score

2

Feb 26, 2026

Kinetics-Sounds (test)

74.8Top-1 Acc

2

Feb 26, 2026

automated audio data curation framework

98.58Precision

1

May 15, 2026

Generated (test)

automated audio data curation framework

89.31Precision

1

May 15, 2026

UrbanSound8K (val)

—

—Primary metric

0

Feb 26, 2026

Page 7 of 7

Popular datasets

AudioSet Speech Commands ESC UrbanSound TUT

Follow for updates

Join SOTA2 Contact