What is Gemini TTS? Text-to-speech with the Gemini API: Turkish support, free tier, pricing, SynthID and model shutdowns
You give Gemini TTS text through the Gemini API or Google AI Studio and get natural speech back. The current models are gemini-3.8-flash-tts (130 languages) and gemini-3.8-flash-lite-tts (101 languages); both include Turkish. The pricing page lists 3.8 TTS standard requests on the free tier as “Free of charge”; per-minute limits are not in the documentation. Paid output is $9 (Flash) and $6 (Flash-Lite) per million tokens, doubling on January 1, 2027. The older 3.1 and 2.5 TTS models shut down on November 17, 2026 at the earliest.
- What it does
- Generates natural speech from text
- Free tier
- Yes; no per-minute limit stated
- Paid
- Flash $9 → $18 (January 1, 2027) /1M tokens
- Turkish
- Supported in both 3.8 models
- Older models
- Shutting down November 17, 2026 at the earliest
What is Gemini TTS for? How do you use it?
You give Gemini TTS a text and a voice choice; the model reads the text as natural speech. You can produce single-speaker narration or a two-speaker dialogue. In the new API the text is read verbatim, style and tone go in a separate speech_metadata field, and a one-shot request now returns a WAV file with headers.
- Sign in at aistudio.google.com with your Google account and create an API key.
- Call the model:
model="gemini-3.8-flash-tts" - Pass the voice and text in the request; save the returned WAV.
Who it helps: developers producing podcasts, training videos and in-app voice narration. The setup steps follow the flow of the official documentation; on-screen labels are not specified in the official sources.
Which Gemini TTS model? Does it support Turkish?
| Model | Languages | Status |
|---|---|---|
| gemini-3.8-flash-tts | 130 languages, including Turkish | GA, September 22, 2026 |
| gemini-3.8-flash-lite-tts | 101 languages, including Turkish | GA, September 22, 2026 |
| gemini-3.1-flash-tts-preview | — | Shuts down November 17, 2026 at the earliest |
| gemini-2.5-flash-preview-tts and -pro-preview-tts | — | Shuts down November 17, 2026 at the earliest |
!Code that depends on the older models needs to move to gemini-3.8-flash-tts or -lite-tts before November 17, 2026. There is no official benchmark for Turkish pronunciation quality; test with your own texts.
Model list and shutdown dates from ai.google.dev (October 14, 2026); check language counts on the official page.
Is Gemini TTS free? Pricing and the 2027 increase
- 3.8 TTS standard requests: “Free of charge”
- Per-minute request count not in the documentation
- Paid tier
- Cheaper, fewer languages
According to the pricing page, 3.8 prices double on January 1, 2027. Prices are in US dollars. There is no separate commercial-use statement for the output: the Additional Terms only say that Google claims no ownership and that the user is responsible for the content.
SynthID watermark and voice cloning
Google's blog says every audio clip generated with Gemini audio models is watermarked with SynthID. The TTS developer pages don't describe this; it appears only on the blog.
Voice cloning in AI Studio is not available in Illinois, Texas, the EEA, the United Kingdom, Switzerland and India (blog footnote); the status for Turkey is unclear. Whether voice design and cloning are available on the free tier is not clear in the official sources.
Cancellation, refund and account deletion terms for the API are not covered in this guide; check Google's current terms.
Frequently asked questions
What is Gemini TTS and what is it for?
Models in the Gemini API and Google AI Studio that generate natural speech from text (gemini-3.8-flash-tts and -lite-tts).
Can Gemini TTS read Turkish?
Yes, Turkish is in the language list of both current models. There is no official benchmark for pronunciation quality.
Is Gemini TTS free?
On the free tier the pricing page lists 3.8 TTS standard requests as “Free of charge”; per-minute request counts are not in the documentation. Paid output for Flash is $9 per million tokens.
Is Gemini TTS pricing going up?
Yes, prices for the 3.8 models double on January 1, 2027: Flash from $9 to $18, Flash-Lite from $6 to $12.
What happens to the older Gemini TTS models?
gemini-3.1-flash-tts-preview and the gemini-2.5 TTS models shut down on November 17, 2026 at the earliest; you need to move to the 3.8 models.
Related guides
Sources
Not clear in the official sources: Per-minute request counts on the free tier · a separate commercial-use statement for the output · SynthID coverage on the TTS developer pages · whether voice design and cloning are available on the free tier · Turkish pronunciation quality · a Turkish interface for AI Studio · voice cloning from Turkey · cancellation, refunds, account deletion. Each is covered in its section along with what you can do.