Google ships two Gemini 3.8 TTS models

Google DeepMind on Tuesday released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, positioning the pair as a voice-design toolkit rather than a fixed set of presets. Flash TTS targets character work and creative direction; Flash-Lite is pitched at high-volume dubbing, audio content and voice agents at lower cost.

The headline capability is generative voice design: developers can prompt new voices from scratch, specifying role, accent and vocal characteristics across more than 100 languages and dialects. Google says the library expands from 30 original voices to 2,000+ production-ready options, including Mexican Spanish, Quebec French and Scots English. Voice replication works from a 30-second sample and requires a verbal consent recording from the voice owner that matches the reference speaker; all output carries SynthID watermarking and C2PA credentials. Both models support line-by-line stage directions, long-form generation, native two-speaker scene staging, and scripted vocal bursts such as <laughs> and backchannel cues like |mhm|.

On benchmarks — all cited by Google — Flash TTS takes the top spot on Hume AI's Voice Design Benchmark at 71.4 and leads accent modeling at 60.8, with Flash TTS and Flash-Lite placing first and second on Hume's Overall Quality Index. Google also claims top positions in blind Voice Arena preference tests in Japanese, Brazilian Portuguese, Vietnamese, Modern Standard Arabic, Mexican Spanish and Hindi.

Both models are available now via the Gemini API and a new audio playground in Google AI Studio, and already power Gemini Notebook and Google Vids. Voice remixing is listed as coming soon. Pricing was not disclosed in the announcement.