Mistral releases a new open-source model for speech generation

Published on March 26, 2026
Mistral's new speech model can run on a smartwatch or a smartphone.