From the source
Google DeepMind is releasing two text-to-speech models and a new API endpoint, bringing studio-grade voice generation to the Gemini API.
- Gemini 3.8 Flash TTS targets creative use with high voice fidelity, regional dialects, and stability across long multi-turn conversations.
- Gemini 3.8 Flash-Lite TTS replaces the preview model for high-throughput production and real-time voice agent cascades at lower cost.
- The release also adds voice design, voice replication with consent verification, and a library of over 150 prebuilt and custom voices.
