Model Family
Whisper
OpenAI's open-weights speech-to-text model, robust to noise and 100+ languages.
Definition
Whisper is OpenAI's encoder-decoder speech model trained on 680k hours of multilingual audio. It transcribes 99 languages, translates to English, and handles real-world noisy audio well. Open weights have made it the de-facto STT baseline.
Common use cases
- Transcription
- Translation
- Voice agents