Google launched the Gemini 3.8 Flash TTS models on September 23, 2026. The release includes two variants, gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts, accessible via the Gemini API and Google AI Studio. These models support up to 130 languages and provide tools for generating synthetic speech with control over style, accent, and emotional tone.
The system offers three tiers for voice selection: 30 prebuilt studio voices, an Extended Voice Library with hundreds of options, and a custom voice design tool. Developers can create voices using natural language prompts, which are then assigned a persistent ID for consistent use. The platform also supports voice replication and both single-speaker and multi-speaker audio generation.
On September 15, 2026, Google also introduced Gemini 3.8 Live and 3.8 Live Extended Thinking models. These are designed for real-time, interactive speech-to-speech dialogue. The Extended Thinking variant incorporates a reasoning layer that allows agents to pause and process information before responding, complementing the narration capabilities of the TTS series.
