AI news ·
Mistral AI Launches Voxtral 3 Voice Models and Agent SDK
Mistral AI released Voxtral 3, a voice model set and SDK that combines transcription, generation, reasoning, and tool use in over 50 languages.

Mistral AI has launched Voxtral 3, a set of voice models and an agent SDK that handle transcription, generation, reasoning, and tool use across more than 50 languages. The release brings speech-to-text, text-to-speech, and agentic capabilities into a single development framework.
With these components available off the shelf, the technical barrier to building voice agents is dropping. The practical focus for teams now shifts toward workflow integration, managing latency, and implementing safety guardrails, rather than assembling the core voice stack from scratch.