Overview
- On Wednesday, Sept. 23, Google began rolling out Gemini 3.8 Flash TTS and Flash‑Lite TTS to developers and followed with Gemini 3.8 Live’s Live Avatar becoming generally available to Gemini Enterprise customers on Sept. 24.
- Flash TTS lets developers create new voices from natural-language prompts and can clone a voice from a roughly 30‑second sample after a required verbal consent recording that Google verifies.
- Google built provenance and detectability into outputs by embedding SynthID watermarks in generated audio and adding C2PA credentials for traceable provenance on replicated voices.
- Developers can access the models through the Gemini API and Google AI Studio and will see integrations into Gemini Notebook and Google Vids, though voice‑replication and some features are restricted in jurisdictions such as the EEA, UK, Switzerland, India, Illinois and Texas and enterprise API access is being phased.
- Independent benchmarks reported top rankings for Gemini 3.8 Flash TTS on Hume AI’s Voice Design benchmark and on Artificial Analysis’s pronunciation test, but buyers are advised to validate latency, cost and integration fit for real-world agent and long‑form audio use cases.