At its annual I/O developer conference, Google chief Sundar Pichai and his team rolled out more than 20 announcements during the keynote, touching nearly every Google product and introducing several new ones.
New AI models. A personal agent that works while you sleep. A redesigned search experience. Smart glasses. A shopping cart that follows you across the internet. A tool that builds operating systems in half a day.
The common thread running through nearly every announcement was Gemini, Google’s AI platform, now in its third major generation.
Here is everything major Google announced, what it does, and what it means for its users.
1. Gemini 3.5 Flash
Gemini 3.5 Flash is the first model in Google’s new 3.5 generation, built to balance frontier-level intelligence with speed and cost efficiency. It outperforms the previous-generation 3.1 Pro on coding, agentic, and multimodal benchmarks, while running at 4x the output speed of comparable frontier models.
The model started rolling out immediately in the Gemini app, AI Mode in Search, the Antigravity 2.0 development platform, and the Gemini API for developers.
2. Gemini Omni — A New Creative AI Model
Gemini Omni is an entirely new model series that fuses Gemini’s deep reasoning capabilities with creative generation. Gemini Omni can simulate complex physical concepts like kinetic energy and gravity, and translate difficult scientific ideas into comprehensible videos.
Omni Flash, the first model in this series, can accept images, audio, video, and text as input and produce video output grounded in real-world understanding. The visuals can be edited through natural conversational language rather than technical prompting.
Omni is rolling out to AI Plus, Pro, and Ultra users via the Gemini app, Google Flow, and YouTube Shorts.
3. Gemini Spark — Google’s Most Ambitious AI Agent Yet
Gemini Spark is arguably the biggest headline from the entire event. Rather than a smarter chatbot, it represents a fundamentally different kind of AI: one that acts on your behalf in the background, without requiring you to stay logged in.
Spark runs on Gemini 3.5 Flash and the Antigravity 2.0 platform, operating on virtual machines hosted on Google Cloud. Because it runs server-side, it can continue executing tasks even after you close your laptop or lock your phone.
Users can teach it new skills over time, and it learns your preferences to become a more accurate reflection of your work style.
4. Google Search — Its Biggest Redesign in Decades
AI Mode in Search is now powered by Gemini 3.5 Flash, and the search interface itself has been redesigned to accommodate how people actually search today, with longer, more conversational queries. The input box now expands dynamically as you type, and AI-powered query suggestions now go further than autocomplete, anticipating the intent behind what you're typing rather than just finishing the phrase.
A major new addition is Information Agents—background agents that monitor the web around the clock across blogs, news sites, social posts, and live data like finance, sports, and shopping, then proactively notifying users when something relevant changes. This feature is coming to AI Pro and Ultra subscribers this summer.
Moreover, Google is giving AI Mode in Search the ability to generate custom mini apps—dashboards and trackers built for specific personal use cases like wedding planning or fitness tracking.
5. Antigravity 2.0 — Google’s Agent-First Developer Platform
Antigravity, Google's AI-powered coding platform, is getting a major 2.0 upgrade with an explicit focus on agentic development. It now runs on Gemini 3.5 Flash, comes with Google Cloud-standard privacy protections, a new standalone desktop app, and a command-line interface for developers who prefer working in Terminal.
To demonstrate its power, Google tasked Antigravity 2.0 with building the core of an operating system from scratch. The platform deployed 93 sub-agents, generated 2.6 billion tokens, and completed the foundational framework in roughly 12 hours—at a cost of under $1,000 in tokens. The Antigravity CLI, SDK, and native voice support are all available globally starting today.
6. Universal Cart — AI-Powered Shopping
Universal Cart is a Gemini-powered shopping cart that follows users across the entire Google ecosystem. You can add items while reading Gmail, watching YouTube, browsing Search, or chatting with Gemini. It works across a range of retailers and platforms, including Nike, Target, Walmart, Ulta Beauty, Sephora, Wayfair, and Shopify.
Once items are in the cart, Google’s AI actively works on your behalf: hunting for discounts, tracking price history, flagging stock alerts, and factoring in Google Wallet payment perks and loyalty data to surface the best deal available. At checkout, you can pay directly through Google Pay or hand off to the specific retailer’s own site.
7. Android XR Smart Glasses
Google and Samsung pulled back the curtain on two smart glasses models, developed in partnership with fashion eyewear makers Gentle Monster and Warby Parker.
The glasses let you hold voice conversations with Gemini hands-free, get navigation guidance, receive real-time audio translation in the original speaker’s voice, see on-screen text translated into your language through the camera, and take photos.
Internally, the hardware is built by Samsung and Qualcomm, with the external design handled by Gentle Monster and Warby Parker. Notably, the glasses can pair with both Android phones and the iPhone.
8. Google Pics — AI Image Generation Inside Workspace
Google Pics enables users to upload existing images and edit them by removing or resizing foreground and background elements. Every generation is watermarked with Google’s SynthID technology.
9. Daily Brief — AI-Powered Morning Roundup
Daily Brief is a built-in Gemini feature that generates a personalised summary of what lies ahead each day. It sifts through your Gmail, Calendar, and Tasks, surfaces what's most urgent, organises it by priority, and suggests next steps.
Unlike Gemini Spark, which takes independent action, Daily Brief is read-only and informational.
10. Ask YouTube
Ask YouTube is an AI-powered conversational search tool that lets users discover videos using natural-language questions instead of traditional keyword searches.
Similar to chatting with an AI assistant, users can ask detailed questions and receive AI-generated summaries, recommendations for long-form videos and Shorts, highlighted clips, and suggested follow-up prompts based on their queries.
11. New Google AI Pricing — Ultra Gets More Affordable
One of the most consumer-friendly moves at the event was a significant repricing of Google’s AI subscription tiers. Google AI Ultra now starts at $100 per month, offering 5x higher usage limits than the AI Pro tier. The previous flagship $250 plan has been reduced to $200 and retains all the same capabilities as before.
12. SynthID and C2PA Expand to Fight AI Misinformation
Google's SynthID watermarking and detection system is expanding well beyond the Gemini app, coming to Search and Chrome in the near term. Alongside SynthID, C2PA Content Credentials support is also being added. Google's tools can now determine whether an image was captured directly by a camera or generated by AI, and whether a camera-original image was subsequently edited with AI tools.
In Chrome, users will be able to right-click on any image and ask Gemini to assess whether it was AI-generated.
13. Android Halo — A Unified Dashboard for AI Agents
Android Halo is an upcoming feature that gives users ongoing, at-a-glance awareness of what their AI agents are doing at any given moment. Rather than opening an app to check on a task, Halo surfaces subtle status updates at the top of your phone screen, without interrupting your workflow.
14. Project Genie + Google Maps
Google unveiled a prototype integration between Project Genie and Google Maps’ Street View, allowing users to transform real-world locations into interactive AI-generated environments. Users can select a place in Google Maps and apply visual styles such as ‘Desert Sands,’ ‘Stone Age,’ or underwater worlds, effectively reimagining familiar locations as alternate realities. They can then prompt Genie to create characters and interactive elements within those environments.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.