Skip to main content
BenchmarkSpeech & Audio AIv2015

LibriSpeech

by Panayotov et al. / Johns Hopkins · open-source · Last verified 2026-03-17

LibriSpeech is the standard English automatic speech recognition (ASR) benchmark derived from LibriVox audiobooks, containing 1,000 hours of read speech at 16kHz. Word Error Rate (WER) on clean and noisy test splits drives competitive progress in ASR research.

https://www.openslr.org/12
B+
B+Good
Adoption: A+Quality: AFreshness: C+Citations: A+Engagement: F

Specifications

License
CC BY 4.0
Pricing
open-source
Capabilities
evaluation, speech-recognition, asr-benchmarking
Integrations
Use Cases
model-evaluation, speech-ai, asr
API Available
No
Evaluated Models
whisper-large-v3, conformer-xl, wav2vec2-large, mms
Metrics
wer-clean, wer-other
Methodology
Two test sets: test-clean (2,620 utterances from clean recordings) and test-other (2,939 utterances from more acoustically challenging speech). WER computed with standard normalization; lower is better.
Last Run
2026-01-25
Tags
asr, speech-recognition, english, audiobooks, wer
Added
2026-03-17
Completeness
100%

Index Score

79
Adoption
94
Quality
88
Freshness
55
Citations
95
Engagement
0

Explore the full AI ecosystem on Agents as a Service