Google's annual developer conference, Google I/O 2026, kicked off with a jam-packed keynote that once again placed artificial intelligence at the center of nearly every announcement. From new AI models and redesigned apps to experimental hardware and shopping tools, the company unveiled a dozen initiatives aimed at reshaping how users interact with search, productivity, and smart devices. Here is a comprehensive breakdown of the 13 biggest announcements from the event.

Gemini 3.5 Flash and Pro

The flagship announcement was the Gemini 3.5 family of AI models. Google launched Gemini 3.5 Flash immediately as the default model for the Gemini app and AI Mode in Search, with Gemini 3.5 Pro following in the coming weeks. The company emphasized that 3.5 Flash is significantly faster and more capable at handling agentic tasks—like executing multi-step instructions—and offers improved coding abilities, including generating richer web UIs and graphics. Enhanced guardrails also aim to reduce both harmful outputs and false positives when filtering safe queries. This update continues Google’s rapid iteration cycle, building on the momentum from Gemini 2.5 introduced earlier this year.

Gemini Omni – A new model family

Beyond the 3.5 upgrades, Google introduced an entirely new model lineage called Gemini Omni. The first variant, Omni Flash, rolls out today in the Gemini app, Google Flow, and YouTube Shorts. Unlike existing text-to-video tools like Veo, Omni Flash can generate video clips from prompts that combine text, photos, video, and audio. Google’s long-term vision is for Omni to “create anything from any input,” blurring the lines between content types. This positions Google to compete directly with OpenAI’s multimodal models and allows creators to generate short-form videos seamlessly within the Google ecosystem.

Gemini Spark – Always-on AI agent

Dubbed “OpenClaw at home” by some observers, Gemini Spark is an always-on AI agent that runs 24/7 on Google Cloud’s virtual machines. It connects to Workspace apps like Docs, Gmail, Sheets, and Slides, as well as third-party services such as Canva and Instacart. Spark can draft emails, create study guides, scan for hidden fees, and even access local files via the Gemini app on macOS. This marks a significant step toward persistent, personalized AI assistants that work in the background without requiring manual prompts.

Build full Android apps with prompts in AI Studio

Google AI Studio now enables users to “vibe-code” entire native Android apps using natural language prompts. The feature includes an embedded Android emulator for live previews, the ability to install apps directly on connected phones, and export options to Android Studio, GitHub, or ZIP files. Google also announced plans to allow publishing apps exclusively for friends and family, and later add Firebase integration. This democratizes app development, lowering the barrier for non-programmers to create functional applications.

Project Aura smart glasses update

Project Aura, Google’s smart glasses collaboration with Xreal, received a hardware refresh. The external compute puck is now slimmer and includes a fingerprint sensor and lanyard for convenience. New features showcased include Gemini-powered widgets for Google Calendar and Keep, as well as improved overall Gemini performance in the XR platform. The glasses are positioned as a more immersive alternative to current audio-only smart glasses, with a form factor that aims for mainstream acceptance.

New Android XR glasses from Warby Parker and Gentle Monster

In addition to Project Aura, Google announced two audio-only Android XR smart glasses models from Warby Parker and Gentle Monster, launching this fall. Announced initially at I/O 2025, these glasses now have finalized designs. They will support live translation, navigation assistance with Gemini, and notification summaries, similar in scope to the Ray-Ban Meta Smart Glasses but integrated directly into Google’s ecosystem. The partnerships leverage Warby Parker’s optical retail network and Gentle Monster’s fashion-forward aesthetics.

Google’s universal cart for one-click checkout

Google introduced a “Universal Cart” that lets users add products from YouTube, Search, Gemini, and Gmail, then check out from multiple merchants simultaneously. Partners include Nike, Target, Walmart, Ulta Beauty, Sephora, Wayfair, and Shopify. The cart automatically flags potential issues, such as incompatible PC parts, and integrates loyalty info from Google Wallet to maximize savings. This feature launches in Search and Gemini this summer, with YouTube and Gmail support arriving later. It represents Google’s most ambitious push yet into unified e-commerce.

Gmail Live – Voice search for your inbox

Gmail is gaining a voice-driven search capability similar to Gemini Live. By tapping an icon in the search bar, users can ask questions orally and receive extracted answers from emails instead of wading through long threads. The feature can quickly pull up confirmation codes, receipts, or other details. Similar voice-based functionality is coming to Google Docs and Keep, with the ability to query data from Drive and Gmail. This makes email search more conversational and efficient.

Google Pics – AI image editing in Workspace

Workspace is getting a new app called Pics, powered by the Nano Banana 2 chip and Gemini. Users can click on any part of an image and leave a comment describing changes, rather than crafting detailed text prompts. This iterative editing approach aims to streamline AI image manipulation for professionals and casual users alike. Google plans to integrate Pics’ capabilities into other Workspace apps over time.

Search gets agents, generative UI, and mini apps

Google Search is evolving significantly. The search box now expands for longer queries, supports multiple input types (text, images, files, videos, Chrome tabs), and offers AI-generated suggestions. New “information agents” provide summarized updates on specific topics by pulling from blogs, news, and social media. A “generative UI” feature creates on-the-fly visuals like simulations and interactive tables. Additionally, Search can generate “mini apps”—custom dashboards for frequently searched tasks, such as event planning. The agent functionality debuts this summer for AI Pro and Ultra subscribers.

AI Ultra subscription price cut

Google slashed the price of its premium AI Ultra subscription, introduced last year at $249.99/month. The new structure starts at $100/month, with a $200/month tier that includes access to Project Genie. This aligns with OpenAI’s pricing and aims to attract more power users to Google’s advanced models, including higher usage limits and early access to experimental tools.

Expanded AI detection tools

To combat misinformation, Google is expanding SynthID watermarking and C2PA Content Credentials to Chrome and Search. Uploading or selecting images via Search, Lens, or Circle to Search will now reveal provenance details. A future Chrome feature will let users circle questionable images on any website to check their origin. This move strengthens Google’s commitment to transparency as AI-generated content proliferates.

AI agents for Google Beam (formerly Project Starline)

Google is experimenting with lifelike AI agents in its Beam video-calling technology. The first demo featured “Sophie,” an agent that appears on screen (though not in 3D), responds to questions, reads documents held up to the camera, and makes recommendations. Google also demonstrated group calls via Beam, compatible with Google Meet and Zoom. This could redefine remote collaboration by blending human presence with intelligent virtual assistants.

Source: The Verge News