Skip to main content
BenchmarkSpeech & Audio AIv17.0

Common Voice

by Mozilla Foundation · open-source · Last verified 2026-03-17

Common Voice is Mozilla's crowd-sourced multilingual speech corpus spanning 100+ languages with verified recordings from volunteers. It benchmarks ASR systems on low-resource and diverse language conditions, making it critical for evaluating cross-lingual speech model generalization.

https://commonvoice.mozilla.org
B+
B+Good
Adoption: AQuality: AFreshness: ACitations: AEngagement: F

Specifications

License
CC0 1.0
Pricing
open-source
Capabilities
evaluation, multilingual-asr, low-resource-evaluation
Integrations
huggingface
Use Cases
model-evaluation, speech-ai, multilingual-asr
API Available
No
Evaluated Models
whisper-large-v3, mms, wav2vec2-large-xlsr
Metrics
wer, cer
Methodology
Models evaluated on the official test split per language. Average WER over languages weighted by test-set size. CER reported for logographic languages. Results typically reported for v15.0 or latest available.
Last Run
2026-02-08
Tags
asr, multilingual, crowdsourced, speech, wer
Added
2026-03-17
Completeness
100%

Index Score

73.5
Adoption
88
Quality
84
Freshness
88
Citations
86
Engagement
0

Explore the full AI ecosystem on Agents as a Service