Three Gemini TTS preview ids shut down Nov 17: what replaces them
gemini-3.1-flash-tts-preview and both 2.5 TTS previews are listed for shutdown on Nov 17, 2026. Google names two replacements; here is the price per 10 seconds.
Google's deprecations page lists three Gemini TTS preview ids for shutdown on November 17, 2026: gemini-3.1-flash-tts-preview, gemini-2.5-flash-preview-tts and gemini-2.5-pro-preview-tts. All three point to gemini-3.8-flash-tts or gemini-3.8-flash-lite-tts. The two replacements went GA on September 22, so you can switch today.
The ids and their price per 10 seconds
Google counts 25 audio tokens per second, so 10 seconds is 250 tokens. The per-10-second figures below are the pricing page's own equivalents on the paid Standard tier, or the same division for the old previews, whose page lists only a per-million price.
| Model id | Shutdown | Audio price per 1M tokens | Per 10 s of audio |
|---|---|---|---|
| gemini-3.1-flash-tts-preview | Nov 17, 2026 | $20.00 | $0.005 |
| gemini-2.5-flash-preview-tts | Nov 17, 2026 | $10.00 | $0.0025 |
| gemini-2.5-pro-preview-tts | Nov 17, 2026 | $20.00 | $0.005 |
| gemini-3.8-flash-tts (to Dec 31, 2026) | GA | $9.00 | $0.00225 |
| gemini-3.8-flash-lite-tts (to Dec 31, 2026) | GA | $6.00 | $0.0015 |
The price catch
The 3.8 prices are promotional until December 31, 2026 and double on January 1, 2027, to $18 and $12 per 1M audio tokens. At those rates Flash is $0.0045 and Lite is $0.003 per 10 seconds. So from January, Lite costs more than the 2.5 Flash preview did at $0.0025. Budget on the January numbers, not the launch ones.
What to change
- Cheap, high-volume narration or voice-agent speech:
gemini-3.8-flash-lite-tts. Google says it was built to replacegemini-3.1-flash-tts-preview. - Expressive or long-form voiceover:
gemini-3.8-flash-tts. - Re-listen to your output before you switch. Both are new models, and the September 22 release note also adds voice design and a
/v1beta/voicesendpoint.
Sources
Related posts
More in Models
- Gemini voice remixing is 'coming soon': what Sume TTS can change today
Google says Gemini voice remixing for timbre, pitch, pace and accent is coming soon. Sume TTS today has speed, volume and an emotion string only.
- Gemini voice replication: the 10-30 second clip and consent checklist
Gemini 3.8 voice replication needs a 10-30 second clean clip and a spoken consent statement. A checklist, and why a Gemini voice name will not work on Sume.
- Google Ads allows ratios between 16:9 and 9:16: Sume models for 3:4
Google Ads lists 16:9, 1:1 and 9:16 and allows ratios between them. Seedance, Wan 3.0 and MiniMax H3 render 3:4 and 4:3 on Sume; Kling 3 and Omni do not.
- Google image model dates on Oct 7, 2026: Nano Banana and Imagen ids
One table of Google's image model release and shutdown dates on Oct 7, 2026, with the replacement Google names for each Nano Banana, preview and Imagen 4 id.
Written by Sume