arxiv:2603.25551

Voxtral TTS

Published on Mar 26

· Submitted by

taesiri on Mar 27

Mistral AI_

Upvote

Authors:

Abstract

Voxtral TTS is a multilingual text-to-speech model that generates natural speech from short reference audio using a hybrid architecture combining semantic token generation and flow-matching for acoustic tokens.

AI-generated summary

We introduce Voxtral TTS, an expressive multilingual text-to-speech model that generates natural speech from as little as 3 seconds of reference audio. Voxtral TTS adopts a hybrid architecture that combines auto-regressive generation of semantic speech tokens with flow-matching for acoustic tokens. These tokens are encoded and decoded with Voxtral Codec, a speech tokenizer trained from scratch with a hybrid VQ-FSQ quantization scheme. In human evaluations conducted by native speakers, Voxtral TTS is preferred for multilingual voice cloning due to its naturalness and expressivity, achieving a 68.4\% win rate over ElevenLabs Flash v2.5. We release the model weights under a CC BY-NC license.

View arXiv page View PDF Project page Add to collection

Community

taesiri

Paper submitter 2 days ago

Voxtral TTS is a multilingual expressive TTS with a hybrid autoregressive semantic token generator and flow-matching acoustic tokens, using Voxtral Codec for high-quality voice cloning from 3 seconds of audio.