Brief IA
›

Mistral AI Launches Voxtral 3 with Transcription at $0.0035 per Minute

Creative AI·Tom Levy·

Mistral AI Launches Voxtral 3 with Transcription at $0.0035 per Minute

Mistral AI Launches Voxtral 3 with Transcription at $0.0035 per Minute
Key Takeaways
1Voxtral 3 combines voice recognition, synthesis, and voice agent SDK
2Hosting available on own infrastructure or via API in the EU
3Transcription available from $0.0035 per minute, supporting over 50 languages
Why it matters : This suite simplifies the creation of multilingual voice agents in real-time while providing assurances on data localization and an attractive entry cost.

Mistral AI offers Voxtral 3, a solution dedicated to real-time speaking conversational agents. The package combines recognition, synthesis, and orchestration, with hosting options in the European Union and a transcription price set at $0.0035 per minute.

EU Hosting and Transcription at $0.0035 per Minute

Mistral AI specifies that Voxtral 3 can be deployed on the client's infrastructure or used via APIs hosted in the European Union, allowing users to maintain control over their data. The transcription rate starts at $0.0035 per minute.

Recognition, Synthesis, and Agent SDK Combined

The Voxtral 3 suite includes a speech recognition model called Voxtral Small Transcribe 3, a text-to-speech model named Voxtral Mini TTS 3, as well as a voice agent SDK based on LiveKit. These components enable the integration of listening, reasoning, tool usage, and vocal output into a single pipeline. Additional information is available in the documentation.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Over 50 Input Languages and 35 Output Languages, with 50 Dialects

Voxtral 3 supports transcription in over 50 languages. Voice generation extends to more than 35 languages and 50 dialects, while incorporating emotional modulation. Agents can interact with third-party applications during real-time exchanges, including scheduling appointments or checking invoices. This system is named Voxtral 3 and is aimed at live voice agent applications.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.