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

Automatic Speech Recognition research benchmarks – Page 34 · SOTA2 Research

Automatic Speech Recognition

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

ESPnet Transformer

0.19WER

3

Feb 26, 2026

ESPnet Transformer

15.5WER

3

Feb 26, 2026

ESPnet Transformer

23.5CER

3

Feb 26, 2026

81.3WER

3

Feb 26, 2026

CHIME5 worn (dev)

47.9WER

3

Feb 26, 2026

CHIME4 et05_real

11.4WER

3

Feb 26, 2026

CHIME4 dt05_real

5.6WER

3

Feb 26, 2026

ESPnet Transformer

4.5CER

3

Feb 26, 2026

ESPnet Transformer

10.6WER

3

Feb 26, 2026

ESPnet Transformer

4.5WER

3

Feb 26, 2026

ESPnet Transformer

6WER

3

Feb 26, 2026

ESPnet Transformer

3.3WER

3

Feb 26, 2026

VoxForge Indian (test)

22.15WER

3

Feb 26, 2026

VoxForge European (test)

0.128WER

3

Feb 26, 2026

VoxForge Commonwealth (test)

8.15WER

3

Feb 26, 2026

VoxForge American-Canadian (test)

4.85WER

3

Feb 26, 2026

23K-word (test)

61.1Frame Accuracy

3

Feb 26, 2026

7WER

2

Jul 8, 2026

FLEURS 30 +10 (test)

43.42WER

2

Jul 7, 2026

FLEURS 30 core12 (test)

31.63WER (%)

2

Jul 7, 2026

ASCEND accented

48.55CER

2

Jul 7, 2026

VoxPopuli SNR = 0 dB (test)

19.51WER

2

Jul 7, 2026

FLEURS SNR = 0 dB (test)

29.42WER

2

Jul 7, 2026

800-hour curated dataset (test)

1.62WER

2

Jul 3, 2026

IWSLT short-form EN-EN 2026

12.3WER

2

Jun 26, 2026

Page 34 of 38

Popular datasets

LibriSpeech FLEURS Common Voice AISHELL Multilingual ASR Corpus

Follow for updates

Join SOTA2 Contact