Mistral AI has released Voxtral TTS, an open-weight text-to-speech model with 4B parameters. It supports 9 languages, adapts to custom voices with just 3-5 seconds of reference audio, and delivers comparable quality to ElevenLabs Flash v2.5 at lower cost. Available via API and open weights on Hugging Face. https://mistral.ai/news/voxtral-tts #AIagent #AI #GenAI #AIInfrastructure #Mistral