PlanAssist expands to 12 more councils in England
A Gemini tool for planning officers, not a consumer product, is entering a 12-month beta after pilots in three councils.
Every Google Gemini update in one feed — release notes, features and changes, straight from the source.
73 updates
A Gemini tool for planning officers, not a consumer product, is entering a 12-month beta after pilots in three councils.
Google's October roundup puts a local coding agent in the Mac app and lets friends share avatars, with age, language, and region limits on several items.
Google's Gemini API moves Computer Use to general availability in both the Interactions and GenerateContent APIs.
Google started the iOS and Android rollout on October 15 for listed paid plans, and only for US English accounts.
Generate a document, PDF, deck, or spreadsheet and revise it in the chat. PDFs work on mobile. The other formats are web only.
Topics such as the Silk Road, a jet engine, and lightning open as layered scenes in the Gemini app, on the web and on mobile.
Google’s latest frontier model ranks first on a new intelligence index, but access starts with cyber defenders in the Fairwind Program.
Developers can register one app for Search, Gemini, or both, though access to this preview opens in phases.
Signed-in Gemini users can ask whether a file carries Google’s SynthID watermark. A missing mark does not prove the file is human-made.
A stateless, zero-dependency command-line client for coding agents, scripts and CI pipelines, built on the Interactions API and now in beta.
Signed-in users can ask Gemini whether an image, video or audio file has Google's SynthID mark, though it only recognizes Google AI content.
New paid-tier projects now log Generate Content API requests by default; existing projects are unchanged, and opting out is per request or in settings.
With your permission, Gemini can draw on your Google Maps searches and saved places alongside Google Calendar to tailor suggestions for your day.
The model has a 1M shared context window and supports long decode for complex coding, agentic, math and science tasks.
Requests now route to Gemini 3.8 Flash automatically, while an older Deep Research agent will shut down on October 23, 2026.
The update improves text rendering and character consistency, while the previous model shuts down on October 29, 2026.
Google adds Adobe, Airtable, Linear, Webflow, and more as app connections you can @ mention in Gemini chat.
Google's new TTS models add studio-grade voices, a replacement for the flash-tts preview, and a 150+ voice library via the Gemini API.
Google Chat adds an Ask Gemini shortcut for searching Workspace data, drafting content and managing tasks in-line.
New Gemini API projects are pointed to 3.5 Flash-Lite or 3.8 Flash, while 2.5 model access is now limited to prior users.
Developers parsing Antigravity's local tool calls must update to PascalCase parameters and line-range file edits in the new September build.
Google's Live API gets two GA audio-to-audio models: a low-latency default and an extended-thinking option for heavier reasoning.
Google's Gemini app comes to Windows 10 and 11 with an Alt + Space shortcut, Gemini Spark, and built-in image and video creation.
Google shows six ways Sheets canvas, powered by Gemini 3.8 Flash, turns spreadsheet data into interactive mini apps that sync back to the sheet.
Gemini can now draft slides from Chat, spreadsheets from Drive, emails from Docs, and briefs from Gmail threads without switching apps.
Google AI plans add voice input, a Google Pics image tool, Sheets canvas mini-apps, a browser-linked Gemini Spark, and free student access.
Google details six ways its Workspace Intelligence layer uses existing files and context to cut prompting in Chat, Drive, Gmail, Docs, Slides, and Sheets.
: Released the next generation of Google's music generation models: lyria-3.5-clip-preview: Optimized for 30-second clips, loops, and rapid…
The Gemini API model now generates full songs with natural vocals and outputs 44.1 kHz stereo audio.
Google released Gemini 3.8 Flash, a significant step up from 3.7 Flash for agentic tasks and software engineering at the same price ($0.75 per million…