Gemini 3.1 Flash TTS 即将退役
Gemini 3.1 Flash TTS is being retired
Google is retiring the Gemini 3.1 Flash TTS preview. The workflows, apps and API integrations that already use it keep working until the shutdown.
Shutdown date
- Gemini 3.1 Flash TTS: may stop working as early as November 17, 2026.
After the shutdown, generations with this model will fail. Audio you already generated stays in your library.
How to migrate
Switch your workflows, apps and API calls to one of Google's recommended replacements before November 17:
- Gemini 3.8 Flash TTS: expressive speech for character voices and dialogue, with the same 30 voices and 24 languages, at about 10% less per character.
- Gemini 3.8 Flash-Lite TTS: the fast, lightweight option for high-volume voiceovers.
Both add inline vocal tags like <laugh>, <sigh> and <short pause> to direct the delivery.
For API integrations, replace model_google-gemini-3-1-flash-tts with model_google-gemini-3-8-flash-tts. Text, voice and language work unchanged. For two-speaker scripts, start the text with a speaker line, like "Maya: Hello!". The 3.8 models run on Google's Gemini API, so they can't use your own Vertex AI credentials.
来源：Scenario 官方更新 · scenario.com