# Mistral describes end-of-turn detection in its Voxtral 3 speech model

> Original title: Transcribing speech is one thing. Knowing when someone is done speaking is another.

- Source: Mistral AI
- Published: 2026-11-06T16:38:00.000Z
- Bad Influence: https://badinfluence.news/items/y5wa187sxxmdci04igrc3dcq4
- Original: https://x.com/MistralAI/status/2118744719870066936

## Summary

Mistral says Voxtral 3's end-of-turn detection identifies when a user has finished speaking. The feature is meant to help developers build more responsive voice interactions.

## Post

Transcribing speech is one thing. Knowing when someone is done speaking is another.

🎙 Voxtral 3’s end-of-turn detection helps developers identify when a user has finished speaking, making it easier to build responsive voice interactions without relying on science alone.

 See it in action ⬇️

[Video](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Famplify_video_thumb%2F2118623280009449472%2Fimg%2FXvi2kHnY-ThT4v3-.jpg&mode=thumb&exp=1794182400&sig=d2f10e37777c54fc)