Gemini TTS
Key facts · verified Oct 31, 2026
Gemini TTS is a Google audio generation model released on 1 Apr 2026. Status: Deprecated - shutting down 17 November 2026; replacement: gemini-3.8-flash-tts. Pricing: $0.50 text in / $10.00 audio out per 1M tokens (2.5 Flash Preview TTS); free tier available. Source: ai.google.dev.
Google's preview text-to-speech models in the Gemini API: gemini-2.5-flash-preview-tts, gemini-2.5-pro-preview-tts and gemini-3.1-flash-tts-preview. They speak 130+ languages with 30 prebuilt voices, handle one or two speakers, and return 24 kHz mono WAV audio. Google is shutting all three down on 17 November 2026; the stable successors gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts were released on 22 September 2026, and Google states voice replication is supported only on those two Gemini 3.8 models.
$0.50 text in / $10.00 audio out per 1M tokens (2.5 Flash Preview TTS); free tier available
Not published
Key Features
Full Specifications
$0.50 text in / $10.00 audio out per 1M tokens (2.5 Flash Preview TTS); free tier available | |
|---|---|
| Identity | |
| Developer | |
| Released | Apr 2026 |
| Status | Deprecated - shutting down 17 November 2026; replacement: gemini-3.8-flash-tts |
| Cost | |
| API price | Per 1M tokens, text in / audio out: $0.50 / $10 (2.5 Flash Preview TTS), $1 / $20 (2.5 Pro Preview TTS and 3.1 Flash TTS Preview) |
| Free tier | Yes — Gemini 2.5 Flash Preview TTS is free of charge |
| Capacity | |
| Max length | 32k tokens of context per session |
| Sample rate | 24 kHz |
| Output formats | WAV (16-bit PCM, mono) |
| Capability | |
| Voice cloning | No — 30 prebuilt voices |
| Languages | 130+ |
| Access | |
| API | Yes — Gemini API |
| Measured quality | |
| LMArena Elo | 1340 (checked Aug 2026) |
| LMArena rank | Rank 3 |