Speaking of Voxtral | Mistral AI

Mistral AI launched Voxtral TTS, its first text-to-speech model, for multilingual, emotionally expressive voice generation.

The 4B model supports 9 languages, low time-to-first-audio, and custom voices from short reference clips.

It also includes emotion control and cross-lingual voice adaptation across dialects like American, British, and French.

Mistral says human evaluations rate Voxtral above ElevenLabs Flash v2.5 for naturalness and close to ElevenLabs v3 for quality.

Voxtral TTS is available in Mistral Studio and targets enterprise voice agents and multilingual audio products.