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-Visual Speech Recognition research benchmarks – Page 3 · SOTA2 Research

Audio-Visual Speech Recognition

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

VisG AV-HuBERT Large

1.38WER

4

Apr 2, 2026

LRS3 433 h 10 dB SNR

1WER

4

Mar 5, 2026

LRS3 All Samples (test)

3.7WER (10 dB)

4

Feb 26, 2026

AVSpeech (1,000 manually filtered samples)

25WER

4

Feb 26, 2026

8.9Error Rate (B)

4

Mar 24, 2026

Chinese-English Chinese part (test)

Donut-Whisper small

4.7CER

4

Feb 26, 2026

Chinese-English English part (test)

Donut-Whisper base

4.33WER

4

Feb 26, 2026

MISP2022 (test)

28.06cpCER

4

Feb 26, 2026

Page 3 of 3

Popular datasets

LRS3 LRS2 LRS-3 MuAViC FLEURS

Follow for updates

Join SOTA2 Contact