Google I/O 2026 Highlights: Gemini 3.5-Flash, Antigravity 2.0, AI overhaul in Google Search announced

Google has announced a new image editing that gives users the precision to make changes to their creative projects. The new app will be coming out this summer.

Woodward confirms Gemini app is getting new agents. Gemini power users can create their own workflows using the Gemini app with the new Gemini Spark agent platform

Gemini Omni model is coming to Gemini app from today for Google's paid subscribers. Woodward calls it ‘Nano Banana for video’

We confirmed how Google's new redesign is coming to Gemini app and Josh Woodward just revealed the new redesign at I/O 2026. He also revealed that Gemini app will come with various new voices that come in a number of dialects.

Google is also combining AI Overviews and AI Mode into a more unified experience.

Users will now be able to move seamlessly between traditional search results, AI-generated answers and follow-up conversations without losing context. According to Google, the deeper users go into conversations, the more relevant links and sources become.

The updated AI search experience is rolling out globally today across desktop and mobile devices.

Google has announced one of the biggest overhauls to Search in years. AI Mode in Google Search is now powered by Gemini 3.5 model while the company also unveiled a new AI redesigned search box for AI search.

Unlike the traditional search bar, Google says the new search box is designed to expand dynamically with users' questions and provide AI-powered suggestions while typing. Google says the feature goes beyond simple autocomplete by helping users refine or expand queries with additional context and nuance.

The upgraded search box also supports text, images, files and videos, allowing users to search across multiple formats at once.

During the keynote, the company revealed a new Google AI Ultra tier priced at $100 per month.Google said the new plan is intended to put Gemini Spark into hands of more users.

The company also announced a price cut for its top-tier Ultra subscription from $250 per month to $200 per month.

The leaks are true, Google has announced Gemini Spark which is a new AI-powered agent designed to handle tasks on behalf of users and work continuously in the background.

Google Vice President Josh Woodward called Spark a "personal AI agent" built to help users manage their digital lives while operating under their direction and control.

Unlike a traditional chatbot, Gemini Spark runs on dedicated virtual machines inside Google Cloud and is designed to continue working even after users close their laptops or switch devices. Google says the system is powered by Gemini 3 and its Antigravity framework, allowing it to handle long-running and multi-step tasks in the background.

During the demo, Google showcased Spark handling a variety of tasks automatically. One example involved collecting information from emails, chats and documents to generate a summary email using a personalised writing style.

Google says Gemini Spark is initially rolling out to trusted testers this week and will become available in beta for Google AI Ultra subscribers next week.

The company also revealed its plans of expanding Spark into Chrome as an "agentic browser" capable of completing tasks across the web.

Google has announced Antigravity 2.0 to take on AI agent coding tools from OpenAI and Anthropic. During the I/O presentation, Google DeepMind engineer Varun Mohan said that millions of developers are already using Antigravity, and the new version significantly expands its AI capabilities, integrations and tooling.

Google says Antigravity 2.0 introduces a standalone desktop application designed specifically around AI agents. The new experience focuses on agent conversations, AI-generated artifacts and multi-agent orchestration, allowing multiple AI agents to work on tasks simultaneously.

At the core of Antigravity 2.0 is what Google calls an upgraded "agent harness", a framework that allows Gemini to carry out complex real-world tasks. New capabilities include sub-agents, hooks and asynchronous task management.

Google also revealed that Antigravity has been further optimized for coding tasks. While Gemini 3.5 Flash is said to be four times faster than competing frontier models, Google says it delivers up to 12 times faster performance inside Antigravity.

The company says Antigravity 2.0 is rolling out globally starting today as an early research preview.

Google has officially unveiled Gemini 3.5 Flash, the first model in Gemini 3.5 series. During the I/O keynote, Sundar Pichai said the company has been focused on improving Gemini around agentic coding, long-horizon tasks and real-world workflows.

Pichai says Gemini 3.5 Flash performs better than Gemini 3.1 Pro across most benchmarks while delivering major gains in coding capabilities.

The second key upgrade is speed. Google says Gemini 3.5 Flash delivers performance comparable to other leading frontier AI models while being significantly faster. According to Pichai, the model can generate output at nearly four times the speed of competing frontier systems.

Google also revealed that Gemini 3.5 Flash has already been used internally through its reimagined agent-first development platform called Antigravity. The company says the model has helped accelerate software development workflows inside Google.

Google's invisible watermark SynthId is now also getting support from OpenAI, Kako and Leylan, Sundar Pichai announces.

Hassabis says Gemini Omni Flash is now available across Google products and more updates about the model will be available later in the year.

Google DeepMind CEO Demis Hassabis has launched the new Gemini Omni model. The new model uses Gemini's world knowledge and reasoning abilities and brings a major improvement in kinetic energy and gravity.

Sundar Pichai has officially confirmed that Gemini powered AI search functionity is also coming to YouTube. The feature is called ‘Ask YouTube’ and will be rolling out to users in the US this year.

Soon you will be able to create and edit Docs using just your voice, thanks to the power of Gemini.

Docs Live is coming to Pro and Ultra subscribers this summer and the same capabilities will soon also be available to other Google apps in the future, Pichai says.

Sundar Pichai says Gemini app users have more than doubled from 400 million around a year back to over 900 millions now

Alphabet CEO Sundar Pichai has taken the stage at Shoreline Amphitheatre at Mountain View, California. Pichai says he himself uses Gemini to make sense of his parents' doctor visits.

Google's live-stream has begun and we are expected to soon see Sundar Pichai on stage explaining all the new upgrades coming to Gemini app along with other related apps.

While we wait for Sundar Pichai's official I/O keynote address to begin, the Gemini redesign has officially begun rolling out. As you can see in the screenshot attached above, the redesign is just what the leaks had suggested with a cleaner interface being adopted by Google featuring a glowing pattern at the bottom and the line ‘The Mic is yours’ at the centre.

We are less than an hour away from Alphabet CEO Sundar Pichai getting up on the stage in the Shoreline Amphitheatre in Mountain View, California. Pichai is expected to unveil some of the biggest announcements Google has made in AI including new Gemini models and a new video generation model.