Back to all news
Technology
Mistral open-sources Voxtral, a top-tier text-to-speech model
Open weights, one GPU, aimed at ElevenLabs. The brief.
The FeaturedDaily DeskVerified March 2026
Mistral released Voxtral, an open-weight TTS that runs on one GPU, in March 2026.
What happened. In March 2026 Mistral released Voxtral, a 4-billion-parameter open-weight text-to-speech model that runs on a single 16GB GPU, clones a voice from a few seconds of audio, and supports nine languages.
The claim. Mistral's own blind tests preferred Voxtral over ElevenLabs Flash v2.5 roughly 63-70% of the time (vendor-run; attribute, don't assert).
Sources
- Mistral AI just released a text-to-speech model it says beats ElevenLabs — and it's giving away the weights for free — VentureBeat, 26 March 2026
- Mistral releases an open-weights 'speaking' AI model with Voxtral TTS — SiliconANGLE, 26 March 2026