Gemini Just Took Over Google I/O 2025—Here’s Everything Announced

Google just wrapped up I/O 2025, and if there's one theme running through the entire keynote, it's this: Gemini is no longer just a chatbot-it's the foundation of Google's AI future. From learning how you work to helping you create, code, research, shop, and browse, Gemini is expanding fast-and becoming way more powerful and personal in the process.

Here's a full rundown of everything Gemini-related that was announced this year.

Gemini 2.5 Flash is the New Default

Google confirmed that Gemini 2.5 Flash-the lightweight, fast, and efficient model-is now the default across Gemini. It balances cost, speed, and quality, making it ideal for everyday users and developers alike.

Gemini Live: Now Free with Camera and Screen Sharing

Let's start with the update that will probably affect the most people right away: Gemini Live, which lets you point your phone at real-world stuff or share your screen and talk through it, is now free for everyone on Android and iOS.

Whether it's fixing a broken appliance, getting style advice, or even practicing for interviews, people are having 5x longer conversations with Gemini Live compared to text. In the coming weeks, Google is adding tighter integration with apps like Google Maps, Calendar, Tasks, and Keep, so Gemini can do things like schedule an event or help you plan dinner just by talking.

Agent Mode turns Gemini into a digital doer

Formerly known as Project Mariner, this new Agent Mode in the Gemini app takes things a step further. It's not just responding to your questions-it can now act on your behalf.

You can teach it a task once, and it'll learn how to do similar ones in the future. So if you're searching for an apartment, it can browse websites like Zillow, adjust filters, pull up listings, and even help schedule a tour. A preview version is coming soon for Gemini app subscribers.

Behind the scenes, Gemini is now using new protocols like Agent2Agent and MCP (Model Context Protocol) so these agents can interact and collaborate across services. Expect deeper Chrome, Search, and app integrations soon.

Imagen 4 and Veo 3: Visuals and Video Get a Major Boost

Inside the Gemini app, you can now access Google's latest generation image and video tools:

• Imagen 4: This is Google's new image generation model, offering better text rendering, speed, and realism. It's already built into the Gemini app for everyone-great for everything from event invites to professional slide decks.

• Veo 3: A video generation model with native audio support-think city sounds, character dialogue, ambient background noise, all generated from text prompts. Veo 3 is available today for Google AI Ultra subscribers in the U.S.

Deep Research Gets Personal

Gemini's Deep Research now lets you upload your own files-including PDFs and images-to combine with public sources and generate deep, personalized reports. A market researcher could compare sales data with industry trends, while a student could upload journal articles for a more nuanced literature review.

And soon, Deep Research will integrate directly with Google Drive and Gmail, making it even easier to work across your personal and public information.

Canvas Gets Upgraded with Gemini 2.5 Pro

The Canvas space in the Gemini app just got a serious upgrade. You can now:

• Build interactive infographics, quizzes, and podcasts in 45 languages.

• Use vibe coding to create real working apps just by describing them.

• Harness the full power of Gemini 2.5 Pro's fast and accurate code generation.

If you've ever wanted to go from idea to execution without needing to know how to code, this is a game-changer.

Gemini in Chrome: Ask Questions While You Browse

Starting May 21, Gemini is coming to Chrome on desktop for Google AI Pro and Ultra subscribers in the U.S. You'll be able to click a Gemini button to:

• Summarize or clarify complex web content

• Ask questions about what you're reading

Later this year, Gemini will also be able to work across multiple tabs and navigate websites for you.

AI Mode in Search: Now Smarter and More Conversational

The new AI Mode in Search is here-and it's a whole new way to Google. You can now ask complex, multi-part queries and have a natural conversation with Gemini right from the Search tab.

This new mode is live in the U.S., powered by Gemini 2.5, and is already changing how people search by encouraging longer, deeper queries.

Interactive Quizzes for Students + Free AI Plans for College Users

Gemini is making studying more interactive. You can now:

• Ask Gemini to create tailored practice quizzes

• Get real-time feedback

• Receive personalized follow-up quizzes based on your performance

This is rolling out worldwide, and students in the U.S., Brazil, Indonesia, Japan, and the UK can now get a free school year of Google AI Pro. More countries are coming soon.

Google AI Pro and Ultra Plans

Two new subscription tiers were introduced to power users:

• Google AI Pro (₹1,950/month): Offers expanded access to Gemini, plus tools like Flow, NotebookLM, and higher usage limits.

• Google AI Ultra (₹24,500/month): The VIP plan for AI pioneers. Includes early access to experimental features, the most powerful models, and premium tools like Agent Mode, Deep Think in Gemini 2.5 Pro, Veo 3 video generation, and much more. Ultra subscribers will also get YouTube Premium.

Ultra users get first dibs on everything-plus a 50% discount for the first 3 months (U.S. only, for now).

Gemini Everywhere

From the Gemini app to Chrome, Search, Gmail, Drive, and even student learning tools-Gemini is now the connective tissue across Google's entire ecosystem. And it's just getting started.

You can try most of these features today at gemini.google.com.