Google launches Gemini 3.8 Flash TTS and Flash-Lite TTS models

Google has introduced two new text-to-speech (TTS) models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, designed to enhance audio generation for creators and enterprises. The Flash TTS model focuses on high-quality creative tasks such as character design for games, audiobooks, and podcasts. It allows users to design unique voices from scratch using natural language prompts and provides granular, line-by-line control over acting cues, pacing, and emotional tone. Additionally, the model can replicate a voice using a 30-second audio sample, provided the speaker gives verbal consent.

The Flash-Lite TTS model is optimized for high-volume, cost-efficient applications like large-scale dubbing and voice agents. Both models support over 100 languages and dialects and offer access to a library of more than 2,000 prepackaged voices. To ensure transparency and prevent misuse, Google is implementing SynthID digital watermarking and C2PA credentials for all generated audio.

In parallel, Google is expanding its AI video capabilities within Google Vids by integrating the Gemini Omni 1.1 Flash model. This update allows users to generate 1080p video scenes, extend existing clips while maintaining visual consistency, and specify clip durations. These tools are being rolled out via the Gemini API, Google AI Studio, and Google Workspace products.