Voxtral 3
Mistral AI launched Voxtral 3, a speech-to-text model, a text-to-speech model and a LiveKit-based voice agent SDK for real-time voice agents.
Voxtral 3 has Voxtral Small Transcribe 3 for transcription in 50+ languages, with diarization, context-biasing and end-of-turn detection. Voxtral Mini TTS 3 adds voices that adapt emotionally, with support for over 35 languages and 50 dialects. The SDK joins listening, reasoning, tool calls and speaking in one streaming pipeline, so agents can act on external systems during a live call.
- Date
- Tuesday 20 October 2026
- Lab
- Mistral AI
- Kind
- product
- Access
- closed API
- Price
- From $0.0035 per minute for transcription
Figures
| Measure | Value | Measured by |
|---|---|---|
| Transcription price | $0.0035 per minute starting price | company |
| Transcription languages | 50+ | company |
Available through Mistral's EU-hosted API or for self-deployment; the pages read did not state weights licensing, so access is listed as API.
Sources
- mistral.ai/news/voxtral-3/
- the-decoder.com/mistral-ai-launches-voxtral-3-a-full-voice-agent-stack-that-listens-thinks
This record was checked against its sources on 20 October 2026. How we check