Google’s annual I/O developer conference once again centered on artificial intelligence, with a sprawling keynote that touched on everything from new AI models to hardware partnerships. The event, streamed live from Shoreline Amphitheatre, showcased Google’s ambition to embed AI deeper into its ecosystem—ranging from search and email to glasses and shopping. Below are the 13 biggest announcements from Google I/O 2026.

Gemini 3.5 Flash and Pro

Google launched the next generation of its flagship AI models, starting with Gemini 3.5 Flash, which immediately became the default model for the Gemini app and AI Mode in Search. According to Google, this model is significantly faster and better at handling agentic tasks—meaning it can plan and execute multi-step operations more reliably. It also offers improved agentic coding capabilities, allowing it to generate richer, more interactive web UIs and graphics. A notable improvement is in guardrails: Gemini 3.5 Flash is less likely to produce harmful content and less likely to mistakenly flag safe queries as unsafe. The larger Gemini 3.5 Pro is scheduled to follow next month, promising even greater reasoning depth.

Gemini Omni: A New Model Family

Alongside Gemini 3.5, Google introduced an entirely new family of models called Gemini Omni. The first variant, Omni Flash, rolled out starting the day of the keynote in the Gemini app, Google Flow, and YouTube Shorts. Unlike Google’s Veo model, which generates video only from text, Omni Flash accepts a wide range of inputs—text, photos, video, and audio—to produce video clips. Google envisions that future Omni models will be able to “create anything from any input,” blurring the line between content creation and consumption.

Gemini Spark: Google’s Ongoing AI Agent

Gemini Spark is an always-on AI agent powered by Gemini 3.5 Flash that runs in the background using virtual machines on Google Cloud, available 24/7. It can connect to Google Workspace apps—Docs, Gmail, Sheets, and Slides—as well as third-party services like Canva and Instacart. Spark can write emails, create study guides, monitor for hidden fees, and perform other proactive tasks. Google plans to expand Spark’s capabilities to access local files through the Gemini app on macOS, making it a persistent digital assistant.

Vibe-Coding Full Android Apps in AI Studio

Google announced that users can now build complete native Android apps using natural language prompts directly in Google AI Studio. The feature includes an embedded Android emulator for previewing and editing apps, and users can plug in their phone to install the app for testing. Apps can be exported to Android Studio, GitHub, or saved as ZIP files. Google also said it will soon allow users to publish these vibe-coded apps exclusively for friends and family, rather than publicly. Support for Firebase integrations is coming later.

Project Aura Smart Glasses Update

Google showed off an updated version of its Project Aura smart glasses, developed in collaboration with Xreal. The external compute puck has been redesigned to include a fingerprint sensor and a lanyard, making it easier to wear both the puck and the glasses simultaneously. New features demonstrated included widgets for display glasses, Gemini integrations with Google Calendar and Google Keep, and improved Gemini performance. The glasses represent Google’s continued push into augmented reality.

Android XR Glasses from Warby Parker and Gentle Monster

Two new pairs of Android XR smart glasses are launching this fall, one from Warby Parker and one from Gentle Monster. These audio-only glasses (no display) follow the Ray-Ban Meta form factor and support live translation, navigation assistance with Gemini, and notification summaries. The partnerships were announced at last year’s I/O, but the actual designs were revealed now.

Universal Cart: One Checkout Across Google

Google introduced a “Universal Cart” that allows users to add products from YouTube, Search, Gemini, and Gmail. The intelligent shopping cart works across merchants like Nike, Target, Walmart, Ulta Beauty, Sephora, Wayfair, and Shopify. Users can add items from different stores and check out in a single transaction. The cart can also flag potential issues, such as incompatible parts for a gaming PC, and interpret loyalty info from Google Wallet to maximize savings. Universal Cart launches in Search and Gemini this summer, with YouTube and Gmail support following later.

Gmail Live: Voice-Powered Search

Google expanded Gmail’s search tools with Gmail Live, a voice-driven interface that extracts and delivers information from your inbox based on natural language questions. Instead of sifting through email lists, users can ask for a confirmation code or specific details, and Gmail Live will provide the answer directly. Similar voice-driven AI features are coming to Google Docs and Keep, pulling data from Drive and Gmail.

Google Workspace: New “Pics” App for AI Image Editing

A new app called Pics, powered by Nano Banana 2 and Gemini, allows users to edit images iteratively by clicking on a part of the image and leaving a comment describing the change. This eliminates the need to write full prompts for each edit. Google plans to incorporate Pics’ capabilities into other Workspace apps.

Search Gets Agents, Generative UI, and Mini Apps

The search box has been redesigned with more space for long queries and AI-generated suggestions. Users can now search using text, images, files, videos, and even Chrome tabs. “Information agents” will provide summarized updates on topics by pulling from blogs, news, and social media, launching this summer for AI Pro and Ultra subscribers. A new “generative UI” feature can create simulations, interactive tables, and graphs in Search, and even generate custom “mini apps” for recurring tasks.

AI Ultra Plan Price Cut

The premium AI Ultra subscription, introduced at I/O 2025 for $249.99 per month, now starts at $100 per month. A $200 per month tier includes access to Google’s Project Genie. The price cut aligns Google with OpenAI’s pricing structure.

AI Detection Tools Expand to Chrome and Search

Google is making it easier to identify AI-generated or altered images by expanding SynthID watermarking and C2PA Content Credentials to Chrome and Search. Users can upload or select online images in Search to see provenance details. Chrome will later allow users to circle questionable images on websites to check their origin.

Google Beam Experiments with Lifelike AI Agents

Building on Project Starline, Google is testing lifelike AI agents for Beam. An early demo featured “Sophie,” a video agent that responds to questions, reads documents held up to the camera, and looks up information like restaurant recommendations. Google also demonstrated group calls using Beam, which will work with tools like Google Meet and Zoom.

Source: The Verge News