Gemini 3 Launches With Redesigned Gemini App and Powerful New Generative Interfaces

Google has officially released Gemini 3 Pro, its latest AI model, now available in preview. While global rollout may take a few days, Google has already begun integrating the model across its ecosystem — starting with the Gemini app and the AI Mode in Search. The biggest upgrade is the introduction of generative interfaces, bringing more visual, interactive, and dynamic AI experiences to users.

New Features in the Gemini App

In a blog post, Josh Woodward, VP of Google Labs, Gemini, and AI Studio, announced major upgrades to the Gemini app and web interface. These improvements will be visible on Android, iOS, and the Gemini website.

1. A Redesigned, Modern Interface

The Gemini app now features a cleaner layout with:

• Easier access to start chats

• A new My Stuff folder to find images, videos, reports, and other generated content

• Integrated Google Shopping Graph for product discovery

2. Generative Interfaces (Generative UI)

This is the most significant new capability. According to Google, generative UI can create immersive visuals, custom web pages, tools, games, or applications instantly — all designed in real-time based on your prompt.

The app will initially test two experimental generative interface types:

A. Visual Layout

A magazine-style layout with photos, sliders, and clickable modules.
Example:
Ask Gemini to plan a trip →
It might show:

• a carousel of trip categories

• a slider to choose travel duration

• clickable interest points

The AI then builds a personalised itinerary without requiring detailed typing.

B. Dynamic View

Using Gemini 3’s agentic coding ability, the AI creates a fully custom interface on the spot.
Example:
Ask “Explain the Van Gogh Gallery with life context,” and Gemini generates an interactive window with clickable artwork, scrolling sections, and visual explanations.

Google says users will initially see only one of these experiments as part of A/B testing.

3. Gemini Agent

A new experimental agent feature that can:

• Handle multi-step tasks

• Connect to Google apps

• Manage Calendar events

• Add reminders

• Organise inboxes and draft email replies

• Make online bookings and appointments

Currently available only on web for Google AI Ultra subscribers in the US.

New Features in AI Mode (Search)

Gemini 3 Pro is also rolling out to AI Mode in Google Search, initially for AI Pro and AI Ultra subscribers in the US.

1. “Thinking” Mode

A new setting in the model menu allows users to select the latest Gemini 3 Pro engine. Google says rate limits will increase soon.

2. Better Reasoning & Smarter Results

AI Mode can now:

• Handle more complex, multi-step questions

• Parse large volumes of web pages

• Extract more accurate, context-rich responses

Google is also adding automatic model selection, which sends tough queries to Gemini 3 Pro, while simpler ones go to faster models.

3. Generative Interfaces in Search

AI Mode can now create real-time visual layouts, interactive tools, and simulations inside search results.

Example:
Searching for mortgage loan options →
Gemini 3 creates a live, interactive loan calculator so users can compare scenarios instantly.

“When the model detects that an interactive tool will help you better understand the topic, it generates and codes it in real time,” Google explained.