AI Mode in Search, New Gemini Features: The Biggest Announcements From Google I/O 2025

Google kicked off its I/O 2025 developer conference today, and as expected, there was an avalanche of news about AI. From new Gemini features to AI Mode coming to all Search users in the US, there is quite a lot to digest today, and you’ll find the main highlights below.

New Gemini features

Gemini 2.5 Flash, the company’s fastest and most efficient model, will become the new default choice starting in June. As for Gemini 2.5 Pro, Google’s most intelligent model announced in March, it’s been updated to perform even better with Deep Think, an experimental reasoning mode for highly complex math and coding tasks. Gemini 2.5 Pro and 2.5 Flash are also adding native audio output for a more natural conversational experience.

Starting today, Google is making Gemini Live with camera and screen sharing available for free on Android and iOS. This allows everyone to point their phone at anything and get instant answers about what they’re seeing. Google also plans to integrate Project Astra’s camera and screen-sharing capabilities into Gemini Live. The company believes this will eventually turn Gemini into a universal AI assistant that can understand the world around you.

Gemini’s image generation capabilities are also getting better with Imagen 4, the latest version of Google’s text-to-image model that can produce pictures with even more photorealistic details and clearer typography. The Deep Research feature that turns Gemini into a personal research assistant is also being improved with the ability to upload your own files to guide research, with Gmail and Google Drive integrations coming soon. It’s also possible to turn Deep Research reports into apps, games, interactive quizzes, web pages, and infographics using Canvas in Gemini.

New Google AI Pro and Google AI Ultra plans

The existing Google AI Premium Plan is being rebranded to Google AI Pro. It’s still priced at $19.99/month, but it’s adding new AI filmmaking capabilities in Flow and early access to Gemini in Chrome for users who use English as their Chrome language on Windows and macOS. Starting tomorrow, Gemini in Chrome will be able to summarize information from any webpage, but Google promises new multitasking and agentic capabilities in the future.

The new Google AI Ultra plan, which is now available in the US for $249.99/month (with a 50% discount for the first 3 months for first-time users), will offer the highest usage limits and access to Google’s most capable models and premium features. That includes the highest usage limits for Deep Research, early access to the new Deep Think reasoning mode, and better video generation capabilities with Google’s new Veo 3 model. Google AI Ultra subscribers will also get access to Project Mariner, an AI agent prototype that can manage up to 10 web-based tasks simultaneously. The plan also includes 30TB of Google Drive storage and YouTube Premium.

AI Mode in Search

Starting today, AI Mode is coming to all Google Search in the US. From a new tab in Search, users use AI mode to ask complex questions using natural language and go further with follow-up questions. In the Google mobile app, AI Mode will be available right from the search bar.

AI Mode is different from the AI Overviews that Google has started showing at the top of search results in some markets. Overall, the new feature turns Google into an answer engine that can dive deeper than a traditional Google search.

“AI Mode is where we’ll first bring Gemini’s frontier capabilities, and it’s also a glimpse of what’s to come. As we get feedback, we’ll graduate many features and capabilities from AI Mode right into the core Search experience,” Google said today.

AI Mode in Search is also getting agentic capabilities with Project Mariner, which is now available in the US to Google AI Ultra subscribers. The experimental AI agent can automate multiple web tasks such as event tickets, restaurant reservations, and local appointments in different browser sessions running on virtual machines. There’s also an AI Mode shopping experience that helps users find inspiration and buy products at the best moment with an upcoming agentic checkout feature.

Google Workspace updates

For Google Workspace users, Gemini will power a new speech translation feature in Google Meet. It will start rolling out today to Google AI Pro and Ultra subscribers in beta, initially in English and Spanish, with more languages coming in the next few weeks.

Google Vids, the new video editing app that launched last year, now lets users turn existing Google Slides into engaging videos. Google’s Imagen 4 text-to-image model will also improve the quality of AI-generated images in Slides, Vids, Docs, and other apps with improved spelling and text rendering for product mockups, marketing materials, and more.