Google adds Gemini 3.8 Flash and Flash‑Lite text‑to‑speech models

Google has added two new text‑to‑speech models to its Gemini lineup, expanding voice generation beyond static presets.

The models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash‑Lite TTS, are described as the most expressive audio generation models to date.

Flash TTS targets deep creative direction and character design, while Flash‑Lite TTS emphasizes high‑volume, cost‑efficient scaling.

Both models are available through Google AI Studio, the Gemini API, Gemini Enterprise, Gemini Notebook, and Google Vids.

This broad access lets creators, developers, and enterprises build richer audio experiences across audiobooks, podcasts, games, and interactive media.

Creative control and custom voices

Flash TTS lets users generate entirely new voices from scratch using simple natural‑language prompts.

Users can direct audio line‑by‑line, adjusting pacing, emotion, dialect shifts, and realistic conversational sounds such as backchanneling.

The model supports more than 100 languages and dialects, enabling regional nuances like Mexican Spanish, Quebec French, and Scots English.

Google provides access to over 2,000 production‑ready voices, forming a base that can be expanded into an infinite library.

With as little as a 30‑second sample, developers can recreate a consistent vocal profile, subject to built‑in consent verification.

Scalable deployment and enterprise use

Flash‑Lite TTS is optimized for large‑scale dubbing and expressive voice agents, delivering fine‑grained tone and pacing control at lower cost.

The model is positioned for high‑volume content creation, allowing enterprises to scale audio output without sacrificing expressiveness.

These additions complement earlier Gemini Audio offerings, including 3.5 Live Translate, 3.5 Transcribe, 3.8 Live, and 3.8 Live Extended Thinking.

Both models embed safety mechanisms such as watermarking to help protect generated audio from misuse.

Google notes that generative AI remains experimental, urging responsible use of the new voice capabilities.

Implications for developers and content creators

The ability to synthesize a high‑energy DJ voice from Melbourne or a super‑tinny robot voice demonstrates the range of personas developers can craft.

Enterprises can use the technology to create a consistent brand ambassador voice across multiple touchpoints.

Integration with Gemini Notebook and Google Vids suggests that audio generation can become a seamless part of existing content pipelines.

By offering both a highly customizable model and a cost‑focused variant, Google addresses divergent needs in the growing synthetic‑voice market.

Fine‑grained line‑by‑line direction gives creators unprecedented control over acting cues, making synthetic dialogue more natural.

Why This Matters: The launch gives developers a unified, controllable voice platform that can power large‑scale audio products while embedding safeguards against misuse.