Mistral Launches Voxtral 3 Voice Stack With State-of-the-Art Speech Models

Mistral AI has released Voxtral 3, a comprehensive voice agent platform combining speech recognition, synthesis, and real-time reasoning in a single integrated system, positioning the company as a direct competitor to ElevenLabs, Deepgram, and OpenAI in the rapidly growing voice AI market. The announcement, made October 20, includes three transcription models, a text-to-speech engine, and a voice agent SDK built on LiveKit technology.