Mistral Launches Voxtral: Advanced TTS in Nine Languages
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.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Mistral Revolutionizes Text-to-Speech with Voxtral
The French startup Mistral, specializing in artificial intelligence, has recently introduced Voxtral TTS, its first text-to-speech model. This innovative model supports nine languages, including German, English, French, and Spanish, and stands out for its compactness, integrating four billion parameters.
Voxtral TTS is notable for its ability to produce realistic and emotionally expressive speech. One of its most impressive features is its capability to adapt to new voices using only three seconds of reference audio.
Performance and Accessibility
In terms of performance, Voxtral TTS exhibits a latency of 70 milliseconds for a typical setup, with a 10-second speech sample and 500 characters. In human comparison tests, Voxtral outperformed ElevenLabs Flash v2.5 in terms of naturalness, although ElevenLabs has since introduced an improved version, v3.
Voxtral TTS is accessible via an API, priced at $0.016 per 1,000 characters. Users can also test the model in Mistral Studio, and it is available in open-weight version on the Hugging Face platform.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.