Skip to main content
Version: 1.35

Voxtral Mini 3B

info

This model is currently offered as a preview. It may be removed without a prior deprecation period.

Voxtral Mini 3B is an audio-capable model from Mistral AI that supports automatic speech recognition (speech-to-text) with multilingual language detection.

Model ID​

  • voxtral-mini-3b

Source​

Hugging Face

Modality​

  • Input: Audio (e.g., .mp3, .mp4, .wav)
  • Output: Text

Endpoints​

note

Make sure to use sufficiently high-quality audio with high enough bit rates. We don't recommend MPEG with the MP2 codec, as it's too low quality and can result in cut-off or excessively long generations. Consider using Whisper if you face quality issues.

Rate limits​

The effective monthly quota for the Free tier is 100 minutes/month.