Google’s new Gemini TTS models can clone a voice from 30 seconds of audio

Google released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, enabling users to create new voices from prompts in over 100 languages or clone a voice from a 30-second audio sample with consent verification. The models support line-by-line delivery, two-speaker scenes, and long-form audio, and they earned top scores on Hume AI’s Voice Design Benchmark and Voice Arena tests. Companies such as Figma, HeyGen, and Adobe Firefly are integrating the new TTS models.