Google released Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, two new text-to-speech models in its Gemini Audio family. Flash TTS scored 71.4 on the Hume AI Voice Design Benchmark. The models add prompt-based voice design, line-by-line performance direction, and voice-replication safeguards, and they’re rolling out via the Gemini API and Google AI Studio with no open weights.
Simon Willison’s Weblog·3 weeks ago·
16
● 9 sources
Google released two Gemini text-to-speech models, gemini-3.8-flash-tts and gemini-3.8-flash-lite-tts, along with a playground and a voice library. Generating 1 minute 18 seconds of audio with Gemini 3.8 Flash TTS took about 20 seconds and cost 2.74 cents. Users can now generate longer multi-character conversations and create custom voices using a 30-second audio sample.
Gemini Audio introduced two new text-to-speech models, Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, for creating more expressive, customizable voice outputs. Gemini 3.8 Flash TTS scored 71.4 on Hume AI’s Voice Design Benchmark. The rollout starts today in Google AI Studio and the Gemini API, expanding line-by-line dialogue control, faster high-volume use, and voice customization options across Gemini products and partners.
YouTube introduced custom feeds that let users describe in a prompt the kinds of videos it should recommend, creating a personalized feed pinned on the home page. The feature uses Google’s Gemini AI model and will roll out across web and mobile starting next month. Users will still keep the main recommendations feed, but can switch to their own custom feed tabs for faster browsing.
Gemini 3.8 text-to-speech models were launched for generating custom character voices and directing scene dialogue in Google’s platforms.
The models are identified as “Gemini 3.8.”
This adds new text-to-speech capability across Google AI Studio, the Gemini API, Gemini Enterprise, Gemini Notebook, and Google Vids.
The Sequence Learning Loop issue described several AI failures where models produced plausible explanations but stalled or failed when integrated with the surrounding invoice-processing or scientific-method execution interfaces.
It highlighted three releases from TypeSafe, Google, and Stanford announced last week, including TypeSafe’s Jev and Google’s two Gemini Live models.
The focus shifts from model capability alone to designing better interfaces—what outputs should look like, when tasks are complete, and which steps agents should reproduce versus call directly.
Every AI story that matters,
in your inbox by 8am.
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the
day in two minutes. Follow companies and topics for alerts, or get the
briefing in Slack. Free, no spam, unsubscribe anytime.
Reading TLDRocket needs no cookies, and the readership counts we rely on come from
our own cookieless analytics. Google Analytics is the exception: it sets cookies and
reports to Google, so it stays switched off until you allow it. You can change your
mind any time from “Cookie settings” in the footer.
Strictly necessary
Session security and form protection (tldrocket-session,
XSRF-TOKEN, 2 hours). The site cannot work without them,
so they need no consent.
Always on
Google Analytics 4 (_ga,
_ga_<id>, up to 2 years). Measures which
stories and sections readers use. Google acts as a third-party processor and may
store the data outside the EU. No advertising, no profiling, no data sold.
