GoKawiilGoogle’s I/O 2026 keynote today was once again full of AI-related announcements including a new family of Gemini 3.5 AI models, new features for Search and Gmail, and updates about its Project Aura smart glasses.
If you weren’t able to tune into the event’s livestream today or follow along with our live blog, you can catch up on everything you missed in our roundup below.
Google launched updated AI models at I/O, starting with Gemini 3.5 Flash, with Gemini 3.5 Pro following next month. Starting today, Gemini 3.5 Flash will be the default model for the Gemini app and AI Mode in Search. Google says the new model is significantly faster, better at handling agentic tasks, offers improved agentic coding capabilities, and generates “richer, more interactive web UIs and graphics.” It also has improved guardrails that, according to Google, will make Gemini 3.5 Flash less likely to generate harmful content and less likely to accidentally flag safe queries as unsafe.
Screenshots of Google’s Gemini app with the Neural Expressive design language. Image: Google
Along with Gemini 3.5 Flash, Google is also launching a redesign for the Gemini app. The new look, which Google calls “neural expressive,” features new animations, pops of color, a new font, and haptic feedback. It’s rolling out starting May 19th on the web and in the Gemini app on Android and iOS.
Along with Gemini 3.5, Google is launching an entirely new family of AI models it calls Gemini Omni. The first, Omni Flash, is rolling out starting today in the Gemini app, Google Flow, and YouTube Shorts. At launch, Omni Flash will be able to generate video clips from prompts that include a variety of inputs including text, photos, video, and audio (unlike Google’s Veo model, which is only text to video). Down the line, 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 keep an eye out for hidden credit card fees for you is Google’s ‘we have OpenClaw at home.’ 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 with access to local files through the Gemini app on macOS.
Google will let users vibe-code entire native Android apps now, which they can eventually even publish to the Play Store straight from AI Studio. The feature includes an embedded Android emulator for viewing and editing app previews. Users can also plug in their phone 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 vibe-coded apps exclusively for friends and family, rather than publicly. Support for Firebase integrations is also coming later.
Google showed off an updated version of its Project Aura smart glasses, which it’s collaborating on with Xreal. The glasses were originally unveiled last year, but The Verge’s Victoria Song got a look at an updated version at I/O. The external compute puck for the glasses has been redesigned, and a fingerprint sensor has been added to it (along with a lanyard so you can wear the puck and the glasses). We also got to try out some new features in Google’s XR platform, including widgets for display glasses, Gemini integrations with Google Calendar and Google Keep, and improved Gemini performance.
... continue reading