# Google launches Gemini 3.8 Flash text‑to‑speech models

> Google made its new high‑fidelity TTS models, Gemini 3.8 Flash and Flash‑Lite, generally available via the Gemini API.

Oossa · 2026-09-29 · https://oossa.com/en/google-launches-gemini-3-8-flash-text-to-speech-models

On September 22, 2026 Google announced that its Gemini 3.8 Flash TTS and Gemini 3.8 Flash‑Lite TTS models are now generally available. The flagship Flash model aims for studio‑grade voice quality, regional accents and stable long‑form output, while the Lite version is cheaper and built for high‑throughput or real‑time voice agents. Both are accessed through the new /v1beta/voices endpoint and can create custom vocal personas or replicate consented voices from a library of 150+ options.

## The facts

- Release date: September 22, 2026
- Models: gemini-3.8-flash-tts (high‑fidelity) and gemini-3.8-flash-lite-tts (cost‑efficient)

## Why it matters

The models let developers add realistic, low‑cost voice output to apps, expanding AI‑driven voice assistants and automated narration.

## Sources & references

1. [ai.google.dev: 17 Şubat 2026](https://ai.google.dev/gemini-api/docs/changelog#02-17-2026) – Google Gemini API, 2026-09-29

Last updated: 2026-09-29