What did Google show at the I/O 2026 developer conference

Let’s look at the key technologies that could change how we use smartphones, the internet, and work tools. 

New Gemini 3.5 Flash
Google has unveiled an updated version of the Gemini 3.5 model. It is now the default model in the Gemini app and in AI-powered search features worldwide. The model operates four times faster than its predecessors and outperforms the previous 3.1 Pro version in most benchmarks.

Gemini Omni
Special attention is given to Gemini Omni, a multimodal model capable of generating and editing video on demand based on text, images, or voice prompts. These features are currently being tested on YouTube, including in the Shorts format.

Gemini Spark Personal Assistant
For subscribers to the Google AI Ultra plan, Google announced Gemini Spark. This is a background AI agent that handles routine tasks: it can independently review emails, create summaries, plan tasks, and interact with over 30 third-party services.

“Generative UI” and Ask YouTube
Thanks to AI, the search bar can now instantly generate interactive interfaces tailored to your query. Additionally, the new Ask YouTube feature allows you to skip watching long videos in full; simply ask the AI which minute explains the specific detail you need, and the player will jump directly to that segment.

Tools for Workspace
You can now simply dictate your thoughts aloud, and Gemini will automatically transform them into a beautifully formatted document in Google Docs. A new standalone app, Google Pics, has also been introduced for editing and generating images, allowing you to translate text directly within images, swap objects in photos, and create shared “canvases” for team collaboration.

Smart Glasses without a screen, powered by Android XR 
Google officially unveiled a prototype of its first smart audio glasses, developed in collaboration with Samsung. Designed by fashion brands Warby Parker and Gentle Monster, they look like ordinary stylish eyewear. They feature no display, with all functionality driven by voice and AI. The release is scheduled for autumn 2026.

Android Halo
Android Halo is a new interface layer for AI agents within Android, making their operations visible and transparent to users in real time.

When an AI agent begins performing actions on your behalf—such as searching for flights, organizing emails, or interacting with other apps—the system does not hide this process in the background. Instead, a sleek, glowing Halo icon appears at the top of the screen.

Universal Shopping Cart
The Universal Cart AI agent operates on a specialized shopping protocol. You can ask Gemini to assemble a single cart containing sneakers from one site, a t-shirt from another, and a jacket from a third. The AI will automatically find promo codes, track discounts, and allow you to pay for everything in one click directly through the Gemini or Gmail interface.

AI Studio Update
Google has also significantly updated AI Studio, its online platform for building apps directly in the browser. You can now create full-fledged Android applications simply by describing your idea in words, without installing software or writing code manually. For those working on the go, Google launched a mobile version of AI Studio: start on your phone and continue on your computer, with everything syncing seamlessly.

Antigravity 2.0 and Managed Agents
For developers, Google introduced an entire ecosystem of new tools.

Antigravity 2.0 is an updated desktop application that allows you to run multiple AI agents simultaneously, automate background tasks, and manage Android, Firebase, and AI Studio from a single location.

For those who prefer working in the terminal, there is now Antigravity CLI—a lightweight command-line interface for instantly creating agents without a graphical user interface.

Another new feature is Managed Agents in the Gemini API: with a single call, you can launch a full-fledged agent that reasons, uses tools, and executes code in an isolated environment.

Google Flow
Google Flow, the company’s creative platform for creating videos and music, received an update at I/O 2026. You can now simply converse with the AI, and it will automatically edit the video, select scenes, and keep character appearances consistent throughout the entire clip. In Flow Music, the same model can now create music videos.