Google unveils AI search & agent tools at I/O 2026

Google unveiled a broad set of artificial intelligence products and updates at its I/O 2026 conference, spanning search, software development, media creation and workplace tools.

At the event, it introduced Gemini 3.5 Flash, a new model now generally available through the Gemini API, Google AI Studio and Android Studio. Google also presented Gemini Omni, a model designed to generate video from a range of inputs, with plans to support more output types over time.

Some of the most immediate changes focused on Google Search. AI Mode, Google's AI search product, now has more than 1 billion monthly users, and Gemini 3.5 Flash is the default model globally. Google also merged AI Overviews and AI Mode into a single AI search experience across desktop and mobile.

A redesigned search box now accepts text, images, files, videos and Chrome tabs. Google is also moving toward what it calls information agents in Search, which monitor topics or tasks in the background and send updates to users.

Shopping was another major focus. Google introduced Universal Cart, a system that lets users add items while browsing Search, Gemini, YouTube and Gmail, then track price changes, stock availability and checkout options. It will roll out first in Search and the Gemini app, with YouTube and Gmail to follow.

Agent push

Alongside consumer features, Google put heavy emphasis on AI agents for developers. It rebranded and expanded its development tools under Google Antigravity, which it described as an agent-first platform, and launched Antigravity 2.0 as a standalone desktop application for managing multiple agents.

Google also introduced Antigravity CLI for terminal users and an Antigravity SDK for developers who want to customise agent behaviour and host systems on their own infrastructure. Another central launch was Managed Agents in the Gemini API, which gives developers access to a remote Linux environment where an agent can browse the web, execute code and manage files in an isolated sandbox.

Google AI Studio also received a series of updates. Developers can now build native Android apps inside the platform, connect to Google Play Console for test-track publishing and export projects directly to Antigravity for local development. Workspace data can also be accessed from applications built within AI Studio.

Google also introduced a new USD $100 AI Ultra subscription aimed at developers, technical leads and advanced users. The plan includes higher usage limits than AI Pro, as well as 20TB of cloud storage.

Gemini app

In the Gemini app, Google launched Daily Brief, an agent that creates a personalised digest using connected Google apps such as Inbox and Calendar data. The feature is being made available to Google AI subscribers in the US who choose to connect those services.

Google also outlined Gemini Spark, a more ambitious personal agent designed to work across a user's digital life and take actions under user direction. Spark remains at an early stage and is being released first to trusted testers, before a wider beta for AI Ultra subscribers in the US.

The Gemini interface has also been redesigned under what Google calls Neural Expressive. Changes include a new visual layout, revised tool menus and updates to Gemini Live, which now opens inline and uses a new speech model.

Creative tools

Media creation products were another major area of expansion. Gemini Omni Flash is rolling out through the Gemini app and Google Flow for paid subscribers, and is also available in YouTube Shorts Remix and YouTube Create for adult users at no cost. The model can edit and remix video conversationally while maintaining character consistency across scenes.

Google Flow gained a new agent for multi-step creative tasks and a tools feature that lets users create custom creative utilities using natural language. Google also introduced Google Pics, an image creation and editing tool built on its Nano Banana model, and said it is initially being released to a limited group of testers.

Workplace software was also part of the update cycle. Gmail's AI Inbox is being expanded to more paid subscribers in the US, with new draft reply and task management functions. Google also announced Docs Live, which allows users to create and edit documents by voice, and new voice-driven note organisation in Keep.

Science and hardware

Beyond office and creative software, Google launched Gemini for Science, a set of experimental research tools covering hypothesis generation, computational discovery and literature analysis. It also introduced Science Skills, a bundle that draws on more than 30 life sciences databases and tools for use on agent-based platforms such as Antigravity.

YouTube search was updated with Ask YouTube, a conversational discovery feature that compiles relevant videos and presents results in a structured response. On hardware, Google said the next step for Android XR would be intelligent eyewear, including audio glasses and display glasses, with the first audio product developed with Gentle Monster, Warby Parker and Samsung.

Google also widened access to its SynthID verification system for AI-generated media. SynthID has been used 50 million times globally, and its verification features are being extended to Search and Chrome. Support for C2PA Content Credentials is also being added to help users assess whether content is original or modified.

More companies, including OpenAI, Kakao and ElevenLabs, are adopting SynthID technology for AI-generated content.