Fact · Languages and accents
117,000hours
Of the 680,000 hours of audio used to train Whisper, 117,000 hours cover 96 languages other than English.
Checked against its source on October 16, 2026
In the source's own words
“Of those 680,000 hours of audio, 117,000 hours cover 96 other languages”
Robust Speech Recognition via Large-Scale Weak Supervision (full text), Alec Radford et al., OpenAI (arXiv), 2022, Introduction.
Open the original (DOI 10.48550/arXiv.2212.04356)- What it measures
- Most of the training audio is English. The non-English share is spread unevenly across the 96 languages.
- What it does not show
- Hours per language are published only as a chart. Whisper supports 99 languages in total, so a few have little or no dedicated recognition data.
Source
Cited on
- Dictation Languages, Accents and Voice TranslationTopic
- Windows Dictation Languages: Voice Typing and Voice AccessGuide
- Transcribe and Translate Audio in Other LanguagesGuide
- How can you dictate in a language your device does not support?Question
- What are the best alternatives to Whisper?Question
- What is NVIDIA Parakeet, and is it better than Whisper?Question