Gemini 3.1 Flash TTS, API id gemini-3.1-flash-tts-preview, is a Google text-to-speech model for low-latency speech generation with steerable prompts and expressive audio tags for narration control. Google's Gemini API model page now calls it a legacy preview model. The deprecations page lists a release date of February 26, 2026 and a shutdown date of November 17, 2026, which it describes as the earliest possible date with exact dates communicated in advance.
Key Facts
| Item | Detail |
|---|---|
| API id | gemini-3.1-flash-tts-preview |
| Input and output | Text in, audio out |
| Token limits | 8,192 input, 16,384 output |
| Supported | Audio generation, Batch API |
| Not supported | Caching, function calling, Live API, structured outputs, thinking, Search grounding, and others listed on the page |
| Status | Legacy preview |
| Earliest shutdown | November 17, 2026 |
| Replacements | gemini-3.8-flash-tts, gemini-3.8-flash-lite-tts |
Features
- Steerable speech: prompts guide delivery, and audio tags control narration style.
- Improved multilinguality: Google says the model offers improved multilinguality. The model page does not list voices or languages, so see Google's text-to-speech guide for details.
- Preview terms: Google says preview models may be used in production but typically have stricter rate limits and at least two weeks' notice before deprecation.
Use Cases
- Narration and voiceover generated from text with controllable style.
- Batch generation of audio through the Batch API.
Limitation: this model is in preview and legacy status, so it is a poor base for new production work. Pricing is not covered on the pages we checked on 2026-10-29.
Migration Steps
- Search for
gemini-3.1-flash-tts-previewin your code. - Switch to
gemini-3.8-flash-ttsorgemini-3.8-flash-lite-tts, both listed as stable, following Google's migration guide. - Re-listen to your standard scripts, since voice and style can change between versions.
- Complete the move before the shutdown date.
FAQ
Is Gemini 3.1 Flash TTS still available?
It is available as a legacy preview, but Google lists November 17, 2026 as the earliest shutdown date.
What replaced it?
Google recommends Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS. See Gemini 3.8 Flash TTS in this catalog.
What are its token limits?
8,192 input tokens and 16,384 output tokens.
Does it support the Live API or function calling?
No, both are listed as unsupported on the model page.
Alternatives
- Gemini 3.8 Flash TTS: the recommended replacement.
- Mistral Voxtral TTS: Mistral's speech synthesis model.
- OpenAI tts-1-hd: OpenAI's hosted speech model.
Conclusion
Move off Gemini 3.1 Flash TTS before its shutdown date. Use the Gemini API models page as the source of truth and browse more options in the models category.
Sources
- Websitehttps://ai.google.dev/gemini-api/docs/models
- Social Mediahttps://x.com/GoogleAI
Comments
No comments yet. Be the first to comment!