All the pieces Google introduced at I/O 2026: Largest improve to Search in 25 years, new Gemini 3.5 Flash and Gemini Omni AI mannequin, redesigned Gemini app, and extra

Google hosted its greatest occasion of the 12 months – I/O 2026 yesterday. On the I/O 2026, Google unveiled a variety of synthetic intelligence bulletins, providing a glimpse into how the corporate sees the way forward for know-how. The occasion targeted closely on AI fashions, AI brokers and smarter experiences throughout Google’s ecosystem. The corporate launched two new AI fashions — Gemini Omni and Gemini 3.5 Flash — alongside upgrades to its AI improvement platform and new agent-based instruments for Search, buying and the Gemini app. On the primary day, Google additionally introduced new AI-powered experiences for merchandise comparable to YouTube, wearable gadgets and Google Pics, displaying how deeply the corporate plans to combine AI into on a regular basis digital experiences. Right here’s a glance again at main bulletins made by Google

Gemini Omni turns into Google’s greatest AI announcement

One of many greatest bulletins on the occasion was Gemini Omni, a brand new multimodal AI mannequin household that mixes Gemini’s reasoning skills with superior content material creation capabilities. Google described Omni as “our new mannequin that may create something from any enter — beginning with video.” The corporate mentioned customers can mix textual content, photos, video and audio inputs to create AI-generated movies grounded in Gemini’s understanding of the actual world. “Omni is our new mannequin that may create something from any enter — beginning with video. With Omni, you may mix photos, audio, video and textual content as enter and generate high-quality movies grounded in Gemini’s real-world information. You too can simply edit your movies by way of dialog,” the corporate mentioned in its official announcement.

The primary mannequin within the household, Gemini Omni Flash, is rolling out globally by way of the Gemini app, Google Move and YouTube Shorts. Google mentioned the mannequin will later assist further output codecs comparable to picture and audio era.Google additionally demonstrated conversational video enhancing, the place customers can edit movies utilizing pure language directions. In accordance with the corporate, Omni maintains scene consistency, physics and character continuity throughout a number of edits. The corporate mentioned all movies created utilizing Gemini Omni will embody SynthID digital watermarks for content material transparency and verification.

Gemini 3.5 Flash turns into default Gemini mannequin

Alongside Omni, Google launched Gemini 3.5 Flash, which is now changing into the default mannequin contained in the Gemini app and AI-powered Search. Google mentioned the mannequin is considerably sooner, higher at coding and extra succesful at agentic duties. The corporate additionally claimed Gemini 3.5 Flash contains stronger guardrails to cut back dangerous outputs and unintended blocking of protected prompts. “Gemini 3.5 Flash delivers intelligence that rivals massive flagship fashions on a number of dimensions, on the speeds you might have come to anticipate from the Flash collection. It’s our strongest agentic and coding mannequin but,” the corporate mentioned.

Google mentioned Gemini 3.5 Professional is at the moment in testing and can launch subsequent month. As introduced by the corporate, 3.5 Flash is now the default mannequin for the Gemini app and AI Mode in Search globally. The Gemini app itself can be receiving a redesign with up to date animations, haptic suggestions and a brand new visible fashion that Google calls “neural expressive.”

Gemini Spark turns into Google’s AI agent

Google additionally unveiled Gemini Spark, an always-on AI agent platform that many studies described as Google’s reply to OpenClaw. Powered by Gemini 3.5 Flash, Spark runs constantly on Google Cloud digital machines and might carry out duties like:

Spark integrates with Gmail, Docs, Sheets and Slides, whereas assist for third-party apps by way of the Mannequin Context Protocol (MCP) can be coming. Google mentioned Spark will ask for consumer permission earlier than performing delicate or high-risk actions. Spark is at the moment rolling out to trusted testers and Google AI Extremely customers within the US.

Google expands Antigravity into full AI agent platform

Google additionally introduced main upgrades to Antigravity, its AI-first improvement platform. The corporate mentioned Antigravity is evolving from a coding assistant right into a full agentic improvement system that may assist construct apps, software program and workflows by way of pure language. Google launched:

Throughout demonstrations, Google confirmed Antigravity creating software program and working system parts utilizing AI brokers. The corporate mentioned the objective is to make “anybody a builder,” even with out conventional programming information.

Common Cart brings AI buying to Google ecosystem

One other main announcement was Common Cart, which Google described as a “really clever buying cart.” Explaining the brand new automotive, Google mentioned: “Common Cart is an clever buying cart and your new hub for buying on Google. It really works throughout retailers and throughout providers, so you may add issues to your cart when you’re looking Search, chatting with Gemini, watching YouTube and even studying your Gmail”.“The second you add a product to your cart, it will get to work within the background — discovering offers and worth drops, providing you with insights on worth historical past and alerting you when an merchandise is again in inventory. All of it runs on our Gemini fashions, so your cart will get even smarter because the fashions enhance.”

The system permits customers so as to add merchandise from Search, Gemini, YouTube and Gmail. Customers can then take a look at throughout a number of shops without delay.Google mentioned Common Cart can:

Retailers together with Nike, Walmart, Sephora, Wayfair and Shopify are a part of the system. Google mentioned Common Cart will start rolling out this summer season inside Search and Gemini.

Google Search will get its greatest improve in 25 years

Google introduced main updates to Search with deeper Gemini integration. The redesign pushes AI deeper into each a part of the expertise, drops customers into interactive outcomes for some queries, and introduces brokers that scan the online in your behalf across the clock.The redesigned field expands as you write longer, messier, extra conversational queries—no extra deciding upfront which mode you need. It accepts textual content, photos, recordsdata, movies and even open Chrome tabs as inputs. AI-powered ideas seem as you sort, which Google says go additional than autocomplete. You may ask a follow-up straight from an AI Overview and slide into an AI Mode dialog with out shedding context. The brand new field is rolling out as we speak wherever AI Mode is accessible.The corporate launched:

“We’re getting into the period of Search brokers, the place you may simply create, customise and handle a number of AI brokers on your many duties, proper in Search. We’re beginning with data brokers. Working within the background, 24/7, these brokers intelligently cause throughout data to seek out precisely what you want at precisely the appropriate second,” the corporate mentioned.Search can now reply extra complicated multimodal questions utilizing textual content, photos and movies. Google additionally launched Gmail Stay, which permits customers to look emails utilizing conversational voice prompts.

Ask YouTube brings AI search inside movies

Google launched Ask YouTube, a brand new AI-powered characteristic for looking out inside YouTube movies. The instrument permits customers to ask questions on video content material and obtain contextual responses generated by Gemini AI. Google can be integrating Gemini Omni into YouTube Shorts and YouTube Create.

AI expands throughout Workspace and apps

Google introduced a number of AI upgrades throughout Workspace merchandise. These embody:

Sensible glasses and Android XR return to focus

Google additionally returned to wearable gadgets with a number of Android XR and sensible glasses bulletins. There shall be two varieties of clever eyewear: audio glasses that supply spoken assist in your ear, and show glasses that present you the data you want, proper whenever you want it. Each allow you to keep hands-free and heads up, and get you assist from Gemini simply by asking. Audio glasses are launching first, coming later this fall.The corporate confirmed new clever eyewear developed alongside Samsung, Warby Parker and Mild Monster. The glasses include options comparable to Stay translation, Audio helpNavigation, Gemini AI integration and Notification summaries. Listed here are a number of of the options coming to audio glasses.

In the course of the occasion, Google additionally up to date Venture Aura sensible glasses developed with Xreal.

SynthID expands throughout Google merchandise

Google introduced broader growth of SynthID, its AI watermarking and content material transparency know-how. Increasing the supply of SynthID that was first launched i 2023, Google mentioned: “Our objective is to make it simpler to study extra in regards to the content material you encounter on-line. That’s why we just lately added SynthID verification for picture, video and audio to the Gemini app. Already, it’s been used 50 million occasions globally, and we’re increasing this verification functionality to Search as we speak and Chrome over the approaching weeks”.

“We’re additionally including verification for C2PA Content material Credentials, to simply verify if content material is an unaltered authentic from a digital camera or if it has been modified, and by what instruments. This characteristic is rolling out within the Gemini app beginning as we speak, and it’ll come to Search and Chrome within the coming months. This builds on options just like the labels on YouTube that determine AI-generated content material and our work with trusted testers on Backstory to make detection instruments sooner and extra dependable,” it added.

Gemini app turns into extra agentic

In the course of the Google I/O 2026 keynote, the corporate revealed that Gemini has crossed greater than 900 million folks throughout 230 international locations and greater than 70 languages. Google has redesigned the Gemini app with a design language that it calls Neural Expressive. The interface now options fluid animations, vibrant colours, new typography and haptic suggestions.Google has additionally built-in the Gemini Stay conversational expertise immediately into Gemini. “Now, you may seamlessly change from typing a fast query to diving deep right into a free-flowing dialog — and again once more — with out lacking a beat. We additionally re-engineered the mic so you may faucet and speak by way of a fancy concept at your personal tempo with out getting minimize off mid-thought. And shortly, we’ll begin providing regional dialects, permitting you to decide on a voice that really resonates with you,” the official announcement mentioned.

Neural Expressive is rolling out globally as we speak throughout the online, Android and iOS for everybody.

Google AI Studio at I/O 2026

At Google I/O 2026, Google introduced main upgrades to Google AI Studio, its AI-powered improvement platform designed to assist customers shortly construct functions utilizing synthetic intelligence. The corporate mentioned AI Studio has turn out to be one of many quickest methods for builders and creators to maneuver from a easy immediate to a working manufacturing app, and the most recent updates are aimed toward increasing what customers can construct throughout Google’s ecosystem. One of many greatest bulletins was direct integration with Google Workspace providers, permitting apps constructed inside AI Studio to work with Google Sheets, Drive and paperwork with out leaving the platform. Google mentioned builders can now create dashboards utilizing Sheets information, construct instruments that organise Drive content material and create functions that work immediately with current crew paperwork and workflows. The corporate additionally introduced assist for exporting initiatives immediately into Google Antigravity for native improvement.

In accordance with Google, dialog historical past, undertaking recordsdata and saved secrets and techniques will transfer along with the export, permitting builders to proceed constructing initiatives with their groups whereas scaling improvement workflows extra simply. Google additionally launched a number of new AI-powered design instruments inside AI Studio. These embody customized asset era powered by Nano Banana, which permits the AI Studio Construct agent to mechanically create photos and interface property whereas constructing apps. The corporate mentioned the characteristic is designed to assist builders create customised interfaces and prototypes with out counting on exterior placeholder photos or separate design software program. One other new characteristic permits customers to edit functions immediately inside preview home windows by drawing on parts, annotating layouts and producing up to date visuals in actual time. Google additionally introduced a brand new cellular model of Google AI Studio, which is now accessible for pre-registration. The cellular app will permit customers to create, edit and preview functions immediately from smartphones.