Whisper Large v3

Permissive

OpenAI · Whisper · Released 2023-11

1.55BAudio

The de facto open standard for speech recognition, with strong multilingual accuracy and no usage restrictions.

Strengths

  • +Apache 2.0, unrestricted use
  • +Strong multilingual transcription and translation
  • +Mature ecosystem (faster-whisper, whisper.cpp, WhisperX)

Limitations

  • -Slower than the large-v3-turbo variant for equivalent hardware
  • -Can hallucinate text during long silences or noisy audio

License

Apache 2.0

use commercially with attribution niceties

Hardware

runs on a good consumer GPU (or Apple Silicon) with quantization

Links

Stats

openai/whisper-large-v3— downloads·— likes

via Hugging Face

Related models