Whisper is an open-source OpenAI speech recognition model, released in 2022, that turns speech into text and can translate it into English.
In plain words
Whisper was trained on 680,000 hours of audio collected from the web, about a third of it not in English. It is an encoder-decoder transformer: one part listens to the audio, the other writes out tokens. A single model handles transcription in many languages, translation into English and spotting which language is spoken. Accuracy varies a lot between languages. OpenAI plans to remove its hosted whisper-1 model from the API in February 2027; the open code remains on GitHub.
A quick example
Run the command-line tool on a Japanese audio file and ask it to translate. Whisper writes out an English version of what was said.
Where you'll meet it
- Transcribing interviews, meetings and podcasts.
- Translating speech in other languages into English.
Related
Sources
- DocsIntroducing Whisper, OpenAI · read 29 Sept 2026
- Docswhisper, OpenAI (GitHub) · read 29 Sept 2026
- PaperRobust Speech Recognition via Large-Scale Weak Supervision, Radford et al. (arXiv) · read 29 Sept 2026
- DocsDeprecations, OpenAI API docs · read 12 Nov 2026
Checked by WhatIsAI against the sources above · 12 Nov 2026. Spotted an error? Report it.