Google I/O 2026: From AI agents to smart glasses, here are the biggest announcements | Technology News

Gemini 3.5 Flash with quicker AI brokers and coding instruments

Google has launched Gemini 3.5, its newest household of AI fashions targeted on combining “frontier intelligence” with superior agentic capabilities. The primary launch, Gemini 3.5 Flash, is designed for advanced multi-step workflows, coding, automation, and real-world AI duties whereas delivering considerably quicker efficiency than rival frontier fashions. Google says the mannequin excels in coding, multimodal reasoning, and long-horizon process execution, with purposes starting from app growth and monetary doc preparation to enterprise automation.

Gemini 3.5 Flash combines frontier intelligence with motion capabilities, delivering outputs as much as 4 occasions quicker than rival AI fashions.

Gemini 3.5 Flash additionally powers new AI experiences throughout the Gemini app, Google Search, and the upcoming Gemini Spark private AI agent. The corporate highlighted stronger security protections, enhanced cyber safeguards, and improved AI transparency. Gemini 3.5 Flash is now out there globally by means of the Gemini app, AI Mode in Search, Google AI Studio, Android Studio, and enterprise platforms, whereas Gemini 3.5 Professional is anticipated subsequent month.

Gemini app will get proactive AI brokers, cinematic video instruments and redesigned interface

Google has introduced a significant improve to the Gemini app at I/O 2026, introducing extra proactive AI-powered options, a redesigned interface, and new agentic capabilities. The replace contains Gemini 3.5 Flash, Google’s newest fast-performing AI mannequin, alongside a refreshed “Neural Expressive” design language with fluid animations, richer visuals, and built-in conversational voice interactions. Google additionally unveiled Gemini Omni, a brand new multimodal AI instrument able to producing and modifying cinematic-quality movies utilizing textual content, picture, and video prompts. One other key addition is Every day Temporary, an AI agent that creates personalised morning summaries utilizing Gmail, Calendar, and different linked apps. Google additionally launched Gemini Spark, a cloud-based 24/7 AI agent that may automate recurring duties, organise workflows, monitor emails, and work together with linked companies like Canva and OpenTable. Most of the new options will initially roll out to Google AI Plus, Professional, and Extremely subscribers, whereas a macOS Gemini desktop app with superior voice options can be launching later this 12 months.

Google introduces AI-powered Search with brokers, smarter looking and customized mini apps

Google has introduced a significant overhaul of Search at I/O 2026, introducing new AI-powered options designed to make Search extra conversational, proactive, and agentic. The corporate is upgrading AI Mode with Gemini 3.5 Flash because the default mannequin globally, alongside what it calls the largest redesign of the Search field in over 25 years. The brand new AI-powered Search field helps textual content, photographs, movies, information, and Chrome tabs whereas providing contextual options and conversational follow-ups. Google can be introducing “Search brokers,” which might repeatedly monitor the online, information, buying, finance, and sports activities updates for customers and ship personalised alerts or suggestions.

(Picture: Google)

Search is moreover gaining agentic reserving instruments for companies, native experiences, and buying duties. Google additionally previewed AI-generated mini apps and dashboards powered by Google Antigravity, enabling Search to construct customized trackers, visible simulations, and interactive instruments on demand. Expanded Private Intelligence assist can even permit customers to securely join Gmail, Pictures, and Calendar for extra context-aware Search experiences throughout almost 200 nations and 98 languages.

Gemini Spark: Google’s 24/7 AI agent for on a regular basis duties

Google has unveiled Gemini Spark, a brand new always-on AI agent designed to maneuver Gemini past answering questions and into actively finishing duties on a consumer’s behalf. Powered by Gemini 3.5 and Google’s Antigravity framework, Spark integrates deeply with Workspace apps like Gmail, Docs, and Slides, permitting it to automate recurring workflows, summarise assembly notes, monitor inboxes, and generate reviews or emails within the background. Google says Spark can proceed working even when a tool is locked or offline. The corporate can be increasing integrations with companies like Canva, OpenTable, and Instacart. Spark will start rolling out to trusted testers and Google AI Extremely subscribers within the US.

Gemini Omni Flash for AI video technology and modifying

Google has launched Gemini Omni, a brand new multimodal AI mannequin household designed to generate and edit content material from textual content, photographs, audio, and video inputs, beginning with video creation. The primary mannequin, Gemini Omni Flash, is rolling out to the Gemini app, Google Move, YouTube Shorts, and YouTube Create. Google says Omni combines Gemini’s reasoning talents with generative media instruments, permitting customers to edit movies conversationally whereas sustaining scene consistency, physics, and character continuity throughout a number of prompts.

The mannequin can generate movies grounded in real-world information, create explainers, apply kinds and results, and mix a number of enter codecs into cohesive outputs. Google can be introducing AI-generated digital avatars primarily based on customers’ voices and likenesses, whereas embedding SynthID watermarks into all AI-generated movies for transparency and verification.

Google expands Antigravity with quicker Gemini 3.5 Flash and AI coding instruments

Google has introduced main developer-focused upgrades centred round Gemini 3.5 Flash, its new high-speed AI mannequin designed for agentic workflows and coding duties. The corporate says the mannequin delivers frontier-level efficiency whereas working 4 occasions quicker than rival frontier AI methods. Alongside the mannequin, Google launched Antigravity 2.0, a standalone desktop app for orchestrating a number of AI brokers, in addition to new command-line and SDK instruments for builders. Google can be launching Managed Brokers within the Gemini API, permitting builders to deploy AI brokers able to reasoning, utilizing instruments, and executing code in remoted environments with a single API name.

Antigravity 2.0 updates purpose to present builders extra highly effective instruments and better management over AI brokers.

The corporate additional expanded Google AI Studio with native Android app technology, Workspace integrations, Play Console publishing assist, and a brand new cell app for fast prototyping on the go. Google additionally introduced a brand new $100-per-month AI Extremely subscription tier with greater utilization limits for Antigravity instruments. To encourage AI growth, the corporate unveiled the Construct with Gemini XPRIZE Hackathon, that includes a $2 million prize pool for builders constructing AI-powered purposes utilizing Gemini fashions and instruments.

Google introduces Common Cart for AI-powered buying

Google has unveiled Common Cart at Google I/O 2026, an AI-powered buying hub designed to make on-line purchases smarter and extra proactive. Powered by Gemini and Google’s Purchasing Graph of over 60 billion product listings, Common Cart works throughout Search, Gemini, YouTube, and Gmail, routinely monitoring offers, value drops, inventory alerts, and compatibility points for merchandise added to the cart. Google says the system may even suggest alternate options and determine hidden financial savings utilizing Google Pockets information like loyalty rewards and cost perks. The corporate additionally expanded its Common Commerce Protocol (UCP) and launched the Agent Funds Protocol (AP2), enabling safe AI-assisted purchases with user-defined spending limits and verification safeguards. Common Cart will start rolling out within the U.S. this summer season.

Google Workspace will get Gemini voice instruments, AI Inbox and new picture editor Google Pics

Google has introduced a significant set of AI-powered upgrades for Workspace at I/O 2026, introducing new voice-driven productiveness instruments, an AI picture editor known as Google Pics, smarter inbox administration, and deeper Gemini integration. Gmail, Docs, and Hold are gaining conversational “Stay” options that permit customers search emails, brainstorm paperwork, and organise notes utilizing pure voice instructions. Google additionally unveiled Google Pics, a brand new AI picture technology and modifying app powered by Nano Banana, providing object modifying, textual content translation inside photographs, collaborative canvases, and direct integration with Slides and Drive.

Gmail’s AI Inbox is increasing with contextual draft replies, surfaced file hyperlinks, and improved process administration instruments. Google moreover confirmed that Gemini Spark, its new 24/7 AI agent, will combine with Workspace apps to automate duties and help with emails, paperwork, and scheduling beneath consumer management.

Every day Temporary: Gemini’s new AI morning assistant

Google has launched Every day Temporary, a brand new Gemini-powered AI agent designed to ship a personalised morning abstract of every little thing customers have to know for the day. As soon as enabled, Every day Temporary works throughout linked apps like Gmail and Calendar to assemble pressing emails, upcoming occasions, reminders, and related follow-up data right into a single easy-to-read briefing. Not like a regular abstract instrument, it actively prioritises duties primarily based on consumer targets and even suggests quick subsequent steps. Customers may also refine suggestions over time utilizing easy suggestions controls. Every day Temporary is rolling out first to Google AI Plus, Professional, and Extremely subscribers within the US.

Google teases Gemini-powered good glasses launching later this 12 months

At Google I/O 2026, Google unveiled its upcoming clever eyewear platform constructed on Android XR, introducing AI-powered good glasses developed in partnership with Samsung, Light Monster, and Warby Parker. The primary merchandise shall be audio-focused good glasses arriving later this fall, with future fashions additionally that includes built-in shows. Powered by Gemini AI, the glasses permit customers to get instructions, ship texts, handle calls, translate speech in actual time, snap and edit photographs, and work together with apps utilizing voice instructions without having to take out their cellphone.

(Picture: Google)

Google says the glasses can perceive what customers are , present contextual details about environment, and help with duties corresponding to meals ordering, navigation, and experience reserving. Options embrace hands-free messaging, dwell translations, camera-based AI queries, and integrations with apps like Uber, DoorDash, and Mondly. The glasses will assist each Android and iOS gadgets as Google expands its imaginative and prescient for always-available AI-powered wearable computing.

Google unveils Docs Stay AI voice-writing instrument at I/O 2026

At Google I/O 2026, Google launched Docs Stay, a brand new AI-powered function for Google Docs that may remodel spoken ideas into structured written content material. Designed as a conversational writing assistant, the instrument lets customers freely dictate concepts whereas AI organizes them into outlines, drafts, tables, and formatted paperwork. With consumer permission, Docs Stay may also pull context from linked Google companies like Gmail, Drive, and Chat, in addition to data from the online, to refine generated textual content. Google positions the function as a quicker, extra pure strategy to brainstorm and draft content material utilizing voice fairly than prompts. Docs Stay will launch this summer season for Google AI Professional and Extremely subscribers.

Google redesigns Gemini with ‘Neural Expressive’ AI interface

Google has unveiled Neural Expressive, a significant redesign of the Gemini expertise targeted on making AI interactions extra dynamic and conversational. The brand new interface introduces fluid animations, vibrant visuals, up to date typography, and haptic suggestions throughout internet, Android, and iOS. Google has additionally built-in Gemini Stay instantly into the principle Gemini expertise, permitting customers to modify seamlessly between typing and pure voice conversations. The up to date microphone system is designed to raised deal with free-flowing speech with out interruptions, whereas future updates will add assist for regional dialects. Gemini responses are additionally changing into extra interactive, with AI-generated visuals, timelines, narrated movies, and dynamic graphics changing conventional text-heavy outputs.

Sew with real-time AI design collaboration

Google has introduced main upgrades to Sew at I/O 2026, turning the AI design instrument right into a real-time collaborative platform powered by the Sew Agent. Customers can now create and refine designs utilizing textual content prompts, voice instructions, present codebases, or design information, making the workflow extra pure and interactive. Sew can immediately generate shareable hyperlinks by means of Google AI Studio, whereas accomplished tasks could be exported to Google Antigravity for backend integration or revealed on to the online utilizing Netlify. Google says the brand new options are aimed toward simplifying fast prototyping and dwell collaborative design.

Managed Brokers in Gemini API simplify AI agent growth

Google has launched Managed Brokers within the Gemini API, permitting builders to launch AI brokers with a single API name utilizing the Antigravity agent powered by Gemini 3.5 Flash. The system runs in remoted Linux cloud environments the place brokers can motive, use instruments, browse the online and execute code. Builders may also create customized brokers utilizing markdown-based AGENTS.md and SKILL.md information as a substitute of advanced orchestration code. Classes stay persistent throughout interactions, preserving information and context. Managed Brokers shall be out there by means of the Gemini API, Google AI Studio and Gemini Enterprise Agent Platform, serving to builders construct scalable production-grade AI brokers quicker and with much less infrastructure overhead.