Gemini 3.8 Flash TTS
A high-end text-to-speech model from Google that supports natural language voice design, line-by-line performance control, two-speaker dialogue orchestration, and over 100 language dialects. It targets creative content production scenarios.
Core Facts
Timeline (last 90 days)
Developers should first run the complete workflow with the new model in a test environment, then gradually replace the production environment configuration.
Migrating to new models mainly involves changes to model IDs, rather than a rewrite of the entire API architecture.
During migration, the Live API's session initialization parameters (such as audio sample rate, voice activity detection, and context window configuration fields) and the TTS request body (such as voice selection, output format, and rate parameters) need to be adjusted.
Google has not yet disclosed pricing details, supported language counts, available regions, or specific audio quality comparison data for these two TTS models.
This dual-track strategy causes the intellectual property of voice styles and debugging costs to accumulate on the Flash side, while the marginal production cost is borne by Flash-Lite.
Voice styles created by users in Flash can be directly reused in Flash-Lite.
Flash-Lite is a lightweight version built on top of Flash, trading some degree of customization freedom for lower costs and higher throughput.
Developers can use Flash-Lite to evaluate costs and Flash to evaluate ceiling performance before deciding on the model selection for the production environment
Gemini 3.8 Flash TTS is designed for in-depth creative direction and character design, supporting line-by-line performance control including performance prompts, pacing, dialect switching, and back-channel response details.
For general users, Flash TTS is entering Gemini Notebook, and Flash-Lite TTS is integrated into Google Vids.
7 more timeline events
All Facts (14)
Gemini 3.8 Flash TTS is designed for in-depth creative direction and character design, supporting line-by-line performance control including performance prompts, pacing, dialect switching, and back-channel response details.
75%UnverifiedDevelopers should first run the complete workflow with the new model in a test environment, then gradually replace the production environment configuration.
50%UnverifiedMigrating to new models mainly involves changes to model IDs, rather than a rewrite of the entire API architecture.
50%UnverifiedDuring migration, the Live API's session initialization parameters (such as audio sample rate, voice activity detection, and context window configuration fields) and the TTS request body (such as voice selection, output format, and rate parameters) need to be adjusted.
50%UnverifiedVoice styles created by users in Flash can be directly reused in Flash-Lite.
50%UnverifiedFlash-Lite is a lightweight version built on top of Flash, trading some degree of customization freedom for lower costs and higher throughput.
50%UnverifiedThis dual-track strategy causes the intellectual property of voice styles and debugging costs to accumulate on the Flash side, while the marginal production cost is borne by Flash-Lite.
50%UnverifiedDevelopers can use Flash-Lite to evaluate costs and Flash to evaluate ceiling performance before deciding on the model selection for the production environment
50%UnverifiedThe voice cloning feature allows users to reconstruct a consistent voice profile from just 30 seconds of audio samples, protected by consent verification, SynthID watermarking, and C2PA credentials.
50%UnverifiedVoice cloning requires users to provide a verbal consent recording from the voice owner, which must match the reference speaker before the voice can be created.
50%UnverifiedBoth models support controlling conversation through non-verbal cues such as <laughs>, <sigh>, and <gasp>, as well as active-listening back-channel markers such as |mhm| and |yeah|.
50%UnverifiedThe new TTS models support customizing character, accent, and timbre characteristics through natural language prompts across more than 100 languages and dialects.
50%UnverifiedThe voice library offers more than 2,000 production-ready voices covering regional variants such as Mexican Spanish, Quebec French, and Scottish English.
50%UnverifiedFor general users, Flash TTS is entering Gemini Notebook, and Flash-Lite TTS is integrated into Google Vids.
50%