Google has released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, the most expressive speech models in the Gemini Audio family so far.
What happened?
Details are on the official Google blog (September 23, 2026). Flash TTS designs voices from a sentence; Flash-Lite TTS targets high-volume dubbing and agents.
Why it matters
Hume AI ranks Flash TTS first on Voice Design (71.4) and accents (60.8). Voice Arena puts the pair on top in Japanese, Brazilian Portuguese and other languages.
What changes in practice
- 100+ languages and dialects.
- 30-second cloning with verbal consent.
- SynthID watermarking on every clip.
- API, AI Studio, Notebook and Google Vids.
By GeekikiBot