The 13 most important announcements at Google I/O 2026

Today’s Google I/O 2026 keynote was once again packed with AI-related announcements including a new family of Gemini 3.5 AI models, new search and Gmail features, and updates about the Project Aura smart glasses.

If you are not able to do so Watch the live broadcast of the event today or Follow along with our live blogYou can catch up on everything you missed in our report below.

Google launched updated AI models at I/O, starting with the Gemini 3.5 Flash, followed by the Gemini 3.5 Pro next month. Starting today, Gemini 3.5 Flash will be the default model for Gemini and AI mode in search. Google says the new model is significantly faster, better at handling proxy tasks, provides improved proxy coding capabilities, and generates “richer, more interactive user interfaces and web graphics.” It’s also improved guardrails that, according to Google, will make Gemini 3.5 Flash less likely to create malicious content and less likely to accidentally mark secure queries as unsafe.

Along with Gemini 3.5 Flash, Google is also launching a redesign of the Gemini app. The new look, which Google calls “Neural Expression,” features new animations, pop-in colors, a new font, and haptic feedback. It will be released starting May 19 on the web and in the Gemini app on Android and iOS.

Along with Gemini 3.5, Google is launching a brand new family of AI models that it calls Gemini Omni. The first, Omni Flash, will roll out starting today in the Gemini app, Google Flow, and YouTube Shorts. At launch, Omni Flash will be able to create videos from prompts that include a variety of inputs including text, images, video, and audio (unlike Google’s Veo model, which only converts text to video). In the future, Google says Omni will be able to “create anything from any input.”

The new always-on AI agent that can write emails, create study guides, and monitor hidden credit card charges for you is Google’s “We Have OpenClaw at Home” agent. Powered by Gemini 3.5 Flash, Gemini Spark runs in the background using virtual machines on Google Cloud 24/7, and can connect to Google Workspace apps, including Docs, Gmail, Sheets, and Slides, as well as third-party apps like Canva and Instacart. Google also plans to expand Spark’s capabilities by accessing local files through the Gemini app on macOS.

Google will now allow users to program fully native Android apps, which they can eventually publish to the Play Store directly from AI Studio. The feature includes a built-in Android emulator to view and edit app previews. Users can also connect their phones to install their app directly from AI Studio for testing. Apps can be exported to Android Studio or GitHub, or saved as a ZIP file.

Google says it will also “soon” allow users to publish encrypted apps exclusively to friends and family, rather than publicly. Support for Firebase integrations will also come later.

Google showed off an updated version of its Project Aura smart glasses, on which it is cooperating with Xreal. The glasses were originally It was unveiled last yearbut EdgeVictoria Song of the band Victoria Song gave a look at an updated version at I/O. The glasses’ external computing dial has been redesigned, and a fingerprint sensor has been added to it (with a lanyard so you can wear the dial and glasses). We also got to try out some new features in Google’s XR platform, including widgets for viewing glasses, Gemini integration with Google Calendar and Google Keep, and improved Gemini performance.

Along with Project Aura, two new pairs of Android XR smart glasses will be released this fall, one from Warby Parker and the other from Gentle Monster. Google has announced partnerships with both brands At I/O last yearAnd now we finally know what those glasses will look like. Similar to the basic Ray-Ban Meta smart glasses, these will be audio-only (no screen). It will support features like live translation, navigation assistance with Gemini, and notification summaries.

Google is launching a “universal shopping cart” to which users can add products from YouTube, Search, Gemini, and Gmail. Google says this “smart shopping cart” works across various merchants and services, such as Nike, Target, Walmart, Ulta Beauty, Sephora, Wayfair, and Shopify. So, you can add products from Nike and Target to your global cart and check out of both at the same time.

According to Google, Universal Cart can detect potential issues with products in your cart as well, such as flagging incompatible parts for your gaming PC. It can also interpret perks and loyalty information from Google Wallet payment methods to help users find savings or earn points. Universal Cart will launch in Search and Gemini this summer, with YouTube and Gmail launching later.

Google is expanding Gmail’s already powerful search tools with a version of the Gemini Live experience for your inbox. Tapping an icon in the search bar lets you ask questions by speaking, and instead of having to wade through a list of relevant email threads, the Gmail Live interface will extract the relevant information and present it at your prompting. This could be a faster way to pull up a specific confirmation code for a hotel reservation, for example. Similar voice-based AI features will also be available in Google Docs and Keep and will be able to pull data from Google Drive and Gmail.

Google Workspace is getting a new app called Pics designed to make it easier to apply AI-generated iterative updates to photos. The Pics app is powered by the Nano Banana 2 and Gemini and allows users to make changes by tapping on a specific part of the photo and leaving a comment describing the edits, rather than having to type out a full prompt listing the requested updates. Eventually, Google plans to integrate photo capabilities into other Workspace apps to further simplify the photo editing process.

Google is changing the search box. It will now be expanded to give users more space for longer queries and include AI-generated suggestions, similar to autocomplete. Users aren’t limited to searching using text either – Google will now let you use text, images, files, videos, and even Chrome tabs as search inputs.

Search also gets “information agents,” who can provide you with brief updates on specific topics or questions, taken from blogs, news, social media posts, and more. These agents will launch this summer, starting with AI Pro and Ultra subscribers.

Google’s new Generative UI feature can create visual elements in search, such as simulations or interactive tables and graphs. Search will also soon be able to create “mini apps” for tasks or topics that users search for multiple times, such as a dedicated dashboard for planning an event.

When it was It was introduced last year at I/O 2025A premium Google AI Ultra subscription costs $249.99 per month for access to more advanced AI models, higher usage limits, and early access to new tools. Google is now offering two new pricing tiers for AI Ultra to match OpenAI. It now starts at $100 per month, while the $200 per month option includes access to it Google Genie Project.

In an effort to make it easier for people to identify and confirm when photos were created or modified by AI, Google is expanding its AI detection tools to Chrome and Search. Starting today, uploading images online or selecting them using search (including tools like Google Lens and Circle to Search) can reveal more details about how and where the image was created using Google SynthID watermarking technology and C2PA content credentials. Add-ons will be made available via Chrome that will allow you to circle questionable images on websites to find out their source later.

after Project Starline was renamed Google Beam last yearGoogle is now experimenting with real-life AI agents that can do more than just talk to you during a call. EdgeThe staff’s Sean Hollister demoed an early demo of Sophie; Beam Video Agent is designed to make interactions with AI feel more engaging. Although Sophie didn’t appear in 3D like she does when she’s talking to someone else using… 25,000 hp dimensional systemthe agent can answer questions, read documents displayed in front of the camera, and search for things like restaurant recommendations. During the demo, Google also demonstrated group calling with Beam, a feature that has been in development for a few years and will work with tools like Google Meet and Zoom.