Google releases Gemini 3.8 Flash TTS and Flash-Lite TTS with voice cloning and 100+ languages
Google announced two new text-to-speech models, Gemini 3.8 Flash TTS and Flash-Lite TTS, rolling out today via the Gemini API and Google AI Studio. Both support over 100 languages and offer line-by-line performance control, two-speaker dialogue, and nonverbal cues like laughter. Flash TTS targets creative work such as games and audiobooks, while Flash-Lite TTS is optimized for high-volume,…
What you can do with it
AI at Work →Narrate product demos with Google Vids
What you get You can create training videos and add voice to social clips.
Creates expressive AI voiceovers and avatar narration with better pacing and flow.
- Who for
- marketers and operations
- Cost
- included in a tool you already have
- Effort
- minutes
Use it for
- Narrate product demos
- Create training videos
- Add voice to social clips
Watch out Requires an eligible Google Workspace or AI Pro/Ultra plan
Not for everyone Ultra plan only
Google Vids in the tool directory · 3 changes covered
Key points from the news
- Google launches Gemini 3.8 Flash TTS and Flash-Lite TTS with 100+ language support and line-level directing
- Flash TTS clones voices from 30-second samples with consent verification; all audio watermarked with SynthID
- The Decoder reports pricing: $0.81/hr Flash TTS, $0.54/hr Flash-Lite TTS until 2027, then doubles
The Decoder adds pricing details not in Google's posts: audio output costs $0.81 per hour for Flash TTS and $0.54 for Flash-Lite TTS through end of 2026, rising to $1.62 and $1.08 on January 1, 2027. The Decoder also reports a high-pitched whine artifact in test generations and notes Google has not listed regional endpoints for the new models. The Next Web and Google's blog posts are consistent with the company's announcement on features, benchmarks, and partners such as Figma, HeyGen, and Wondercraft.
Model pages: Gemini 3.8 Flash TTS → · Gemini 3.8 Flash Lite TTS → · Gemini 3.8 Live →
The story so far
4 episodes →- Google releases Gemini 3.8 Flash TTS and Flash-Lite TTS with voice cloning and 100+ languagesthis story
Gemini 3.8 text-to-speech says hello
Google DeepMind · 23 September 2026
Loading the full article…
This text was published by Google DeepMind and written by Leland Rechis. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
14sources- Hacker News discussion · 313 pointsnews.ycombinator.com
- Create more natural, expressive AI voiceovers in Google Vids with upgraded Gemini 3.8 Flash Lite TTSPress · Google Workspace Updates ·
- From "Choosing a Voice" to "Designing a Voice": An AI Engineer's Look at Gemini 3.8 Flash TTSPress · note.com ·
- Gemini 3.8 TTS is amazing. How to 'create a voice' in Google AI StudioPress · note.com ·
- Google's Gemini 3.8 Flash TTS and ChatGPT Voice have become my language teachersPress · note.com ·
- Google launches Gemini 3.8 Flash TTS voice modelsPress · AI News ·
- The Strategic Turning Point for Generative AI Companies in Autumn 2026: Insights from Gemini 3.8 Flash TTS, Claude Opus 5.5, and the Gemini 4 TeaserPress · note.com ·
- Google Releases Gemini 3.8 Flash TTS and Flash-Lite TTS With Prompt-Based Voice DesignPress · MarkTechPost ·
- Google Gemini Unveils Two AI Models for Smarter Voice InteractionsPress · Small Business Trends ·
- Google's new Flash TTS models let you design AI voices from scratch using text descriptionsPress · The Decoder ·
- Gemini 3.8 TTS PlaygroundNewsletter · Simon Willison ·
- Google releases Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, its "most expressive audio generation models yet", with support for more than 100 languagesPrimary source · blog.google ·
- Google’s new Gemini TTS models can clone a voice from 30 seconds of audioPress · thenextweb.com ·
- Google Rolls Out Gemini 3.8 Speech Models In API And AI StudioPress · Unite.AI ·
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- OpenAI faces concerns over loop transformers; Microsoft confirms GPT-6.1 Sol uses them · 1 src
- Meta introduces Muse Wonder and Muse Wonder Full models · 8 src
- Anthropic releases Claude Haiku 5.5 at lower prices than GPT-6 Luna · 1 src
- Microsoft releases Microsoft-Decision-1, a Qwen3.5-9B-based scoring model · 2 src
- Mistral releases Mistral Large 4, nicknamed Le Chonk, a 1-trillion-parameter multimodal model · 1 src
Comments
via GitHub Discussions