April 23.1, 23.1. ·. Our latest audio model features detailed audio tags, giving you fine-grained control to guide AI voices for highly expressive sound generation. — Vilobh Meshram Senior Product Manager. Max Gubin. Principal Research Engineer, representing the Gemini team. Overall summary. Gemini 3.1 Flash TTS has arrived, delivering enhanced AI speech quality and greater control. Audio tags can now be used to modify vocal style and pacing across more than 70 languages.
Google DeepMind News