Google Launches Gemini 3 and Redesigned Gemini App

Brett Tabkeon

Updated 11:55a 11/18:

They have also released Gemini 3 for Search.

Google has introduced Gemini 3 into Google Search’s AI Mode, making it the first time a Gemini model is embedded in Search from day one. The update emphasizes enhanced reasoning, deeper multimodal understanding and stronger agent-capabilities, enabling the system to handle more complex questions and deliver richer output. The model identifies user intent more precisely, expands the search’s “query fan-out” to unearth relevant, credible content it might previously have missed.

A second major change is the rollout of a new generative user interface within Search. Gemini 3 can create customized visual layouts, such as images, tables or interactive tools, to match the nature of the question. For instance, if you’re exploring a physics concept you might see a simulation you can adjust, or when comparing loans you might get a built-in calculator. This shift makes Search less about static links and more about dynamic responses tailored to your query.

Google launches Gemini 3.

Key Highlights

  • Gemini 3 is now live globally
  • Gemini 3 Pro available immediately; toggle via model picker in the Gemini app
  • Higher usage for Google AI Plus subscribers

Core Model Improvements

  • Reasoning and multimodal performance (txt, images, code)
  • "vibe coding" - stronger code generation and Canvas integration
  • More concise, well-formatted, and helpful responses by default
  • Deep Research mode for extended web browsing and analysis

Media Reaction

"Google also redesigned the Gemini app to return answers reminiscent of a full-fledged website, a further blow to content publishers who rely on web traffic to generate revenue." -Reuters

New & Experimental Features (rolling out today)

  • Gemini Agent (US web, Google AI Ultra only initially) Proactive multi-step orchestration across Gmail, Calendar, Tasks, and Workspace - Uses Deep Research, Canvas, and user confirmations for safety
  • Visual Layout - Magazine-style immersive responses with interactive modules and rich media
  • Dynamic View - Agentic UI generation (eg, custom interactive dashboards, trip planners, galleries)
  • "My Stuff" folder - Centralized storage for all generated images, videos, docs, and chats
  • Shopping Graph integration - 50+ billion product listings for instant price comparisons and recommendations

Availability & Rollout

  • Gemini 3 Pro: Global, now
  • Gemini Agent: US web only, Google AI Ultra subscribers first
  • App redesign: Live on iOS, Android, and web

Google positions Gemini 3 as the foundation for the next wave of agentic and generative interfaces, with more features and broader Agent access coming in the following weeks.

Core Model Improvements

Earlier:
Multiple posts, reports, and speculation on the socials this last week hinting that Google has stealth launched Gemini 3. Tiktoks started popping off late on Friday that Mobile Gemini users were seeing a new version of Gemini. Then last night he dropped this on Twitter:

Then, this morning Gemini produced some content almost instantly - no delay at all - and no iterations needed. Clearly some tweaks have been made.