Gemini 3 brings upgraded smarts and new capabilities to the Gemini app

Google has rolled out Gemini 3 inside the Gemini app, positioning it as a major product upgrade rather than a quiet model refresh. In the company’s announcement, Google says Gemini 3 improves reasoning, formatting, and multimodal understanding, while also serving as its strongest “vibe coding” model yet for app-building workflows in Canvas. Gemini 3 Pro is rolling out globally starting now, with paying Google AI subscribers getting higher usage limits and U.S. college students continuing to receive a free year of Google AI Pro.

The product story goes beyond raw model quality. Google is using Gemini 3 to introduce what it calls “generative interfaces,” where the model dynamically creates more visual, interactive response formats instead of returning plain text. The first two experiments are visual layout, which turns prompts into magazine-style interactive views, and dynamic view, which generates custom interfaces in real time for specific requests. Google is also redesigning the Gemini app itself, adding a cleaner layout, easier chat starts, a new “My Stuff” folder for generated assets, and a more integrated shopping experience powered by Google’s Shopping Graph.

The most ambitious piece is Gemini Agent, an experimental feature for Google AI Ultra subscribers in the U.S. that can orchestrate multi-step tasks across tools like Gmail, Calendar, Deep Research, Canvas, and live web browsing. Google says the agent asks for confirmation before critical actions like purchases or sending messages, which makes the launch feel like both a product update and a controlled rollout of a more agentic assistant model. Taken together, Gemini 3 looks less like a single model release and more like Google trying to redefine the Gemini app as a proactive operating layer for search, productivity, and action.

Source: Google official announcement — https://blog.google/products-and-platforms/products/gemini/gemini-3-gemini-app/