Mistral AI has announced the release of Voxtral 3, a new family of models designed for speech-to-text and text-to-speech applications. This suite includes three transcription models and one text-to-speech model, along with a voice agent SDK based on LiveKit.
Among the highlights, the Voxtral Small Transcribe 3 model achieved an impressive 2.1% AA-WER, placing it third on the Artificial Analysis leaderboard. This performance underscores Mistral AI’s commitment to enhancing the accuracy and efficiency of voice technology.
The introduction of Voxtral 3 is expected to significantly benefit developers and businesses looking to integrate advanced voice capabilities into their applications, providing them with the tools necessary for improved user interaction and engagement.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
