Breaking

Google DeepMind Announces Gemini 3.8 Flash TTS and Flash-Lite TTS for Custom Voice Deployment

September 23, 2026 at 11:26 EDT

Media Generation
Foundation Models
AI Agents

Google DeepMind has announced two new text-to-speech models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, aimed at creating and deploying custom audio with distinct voice styles. The company described Gemini 3.8 Flash TTS as a model for designing unique voices with accents and characteristics, while Gemini 3.8 Flash-Lite TTS is positioned for efficiency and scale, letting users choose from created styles or a larger voice set.

September 2026 · Google DeepMind

Gemini 3.8 moves into custom voice

Google DeepMind announced Gemini 3.8 Flash TTS for distinctive custom voices and Gemini 3.8 Flash-Lite TTS for efficient, scaled audio deployment.

new TTS models

input context

64K

output limit

73.7%

DeepSWE v1.1

The two-model TTS split

Gemini 3.8 Flash TTS

Custom voice design with distinct accents and characteristics.

→

Gemini 3.8 Flash-Lite TTS

Efficiency-focused deployment using created styles or a larger voice set.

Benchmark backdrop

Published Gemini 3.8 Flash scores frame the broader model family, not direct TTS quality.

73.7%DeepSWE

89.4%Terminal-bench 2.1

61.4%Finance Agent

54.9%HLE-Verified

Why developers care

Custom voices can support branded agents, avatars, learning tools, accessibility features, and synthetic dialogue at scale.

What remains unclear

Full endpoint names, pricing, release stage, regions, and direct TTS quality benchmarks have not yet been published.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.