Google I/O 2026 highlights: An AI overhaul in Google Search, Gemini 3.5-Flash, Antigravity 2.0, Android XR smart glasses announced

Google had earlier announced that the Android XR smart glasses would arrive in 2026, but the showcase at Google I/O offered the clearest look yet at both the design and the features of the upcoming wearable device.

Google has introduced a new image editing tool that allows users to make highly precise adjustments to their creative work. The app is expected to launch this summer.

The Gemini app is set to receive new AI agents. With the new Gemini Spark agent platform, advanced users will be able to design and customize their own workflows directly within the app.

Google has introduced a new AI Ultra plan priced at $100 per month, while also reducing the cost of its top-tier Ultra plan from $250 to $200.

Google has revealed one of its biggest Search updates in years. AI Mode is now powered by the Gemini 3.5 model, alongside a newly redesigned AI-driven search box. Unlike the classic search bar, the new input field expands as users type and offers intelligent suggestions in real time. Google says it goes beyond basic autocomplete by helping users shape, refine, and deepen their queries with added context.

Google Vice President Josh Woodward described Spark as a “personal AI agent” designed to help people organize and manage their digital activities while staying fully under user control. Unlike conventional chatbots, Gemini Spark operates on dedicated virtual machines within Google Cloud and can keep running even when users shut their laptops or change devices. Google said it is powered by Gemini 3 along with the Antigravity framework, enabling it to perform complex, long-duration, multi-step tasks in the background.

Google has introduced Gemini 3.5 Flash, marking the first release in the Gemini 3.5 lineup. Speaking at the I/O keynote, Sundar Pichai highlighted that the focus has been on enhancing agent-style coding, extended multi-step tasks, and practical real-world workflows. He added that Gemini 3.5 Flash outperforms Gemini 3.1 Pro on most benchmarks, showing notable improvements in programming and development tasks.

A major highlight is its speed upgrade. Google claims Gemini 3.5 Flash matches the performance of other top-tier frontier AI models while running much faster. Pichai noted that it can produce responses at almost four times the speed of comparable leading systems.

Google DeepMind CEO Demis Hassabis has introduced the new Gemini Omni model. He said it builds on Gemini’s broad knowledge and reasoning capabilities, delivering significant advancements in understanding motion-related physics such as kinetic energy and gravity.

Sundar Pichai has announced that YouTube will soon get a Gemini-powered AI search feature. Called “Ask YouTube,” it will begin rolling out to users in the United States later this year.

Users will soon be able to build and modify documents entirely through voice commands, powered by Gemini. Pichai also said the “Docs Live” feature will roll out this summer for Pro and Ultra subscribers, with similar functionality planned for other Google apps in the future.

Sundar Pichai noted that the Gemini app has seen rapid growth, with its user base rising from about 400 million roughly a year ago to more than 900 million today.

Alphabet CEO Sundar Pichai has begun his keynote at the Shoreline Amphitheatre in Mountain View, California. During his address, he shared that he personally relies on Gemini to help interpret and organize information from his parents’ medical appointments.

Google’s livestream has started, and Sundar Pichai is expected to take the stage soon to unveil new updates for the Gemini app and other related services.

While we wait for Sundar Pichai’s official Google I/O keynote to begin, the redesigned Gemini experience has already started rolling out. As leaked earlier, the update brings a cleaner interface across Google’s design language, highlighted by a glowing pattern at the bottom and the centered tagline, “The mic is yours.”

Rumors indicate that Google may introduce a new video generation system called Gemini Omni at I/O 2026. The feature was reportedly spotted in limited testing within the Gemini app, where some users saw a prompt describing it as a tool to “remix videos, edit them directly in chat, use templates, and more.” Early clips shared online suggest the model may be capable of producing highly realistic video outputs, including detailed scenes with text elements and human-like interactions. Some reports also hint that Omni might not be a standalone system, but rather an extension or evolution of Google’s existing Veo video generation technology.

Google may be gearing up for a stronger move into agent-based AI with a new capability reportedly called “Gemini Spark,” as per a 9to5Google report. The feature is said to be spotted in early testing within a redesigned Gemini app, which may introduce a split interface featuring separate “Chat” and “Agent” sections.

Unlike today’s chatbot-style interaction, Spark is expected to go beyond simple responses and actually perform tasks for users. Early leaks indicate it could handle actions like organizing email inboxes by summarizing or archiving promotional messages and unsubscribing from newsletters, generating meeting summaries with relevant context, and building customized news briefings that update over time.

We are expecting a wave of announcements from Google across Cloud, Gemini, Workspace, Search, and possibly even its XR divisions today, with the keynote likely to run close to two hours.

Similar to this year, Google hosted the Android Show separately last year, giving the company additional time to highlight its latest advancements in artificial intelligence. Among the key announcements were Gemini 2.5 Pro, the Veo 3 video generation model, Imagen 4 for AI-powered image creation, and Project Moohan, which focuses on Android XR smart glasses.

Google also revamped Search with a dedicated “AI Mode” powered by Gemini 2.5, making results more conversational and intelligent. For developers, Google introduced Jules, an AI pair programmer integrated directly into IDEs to help detect bugs and streamline code refactoring.

Google is expected to showcase its Android XR-powered smart glasses at Google I/O 2026, marking a major step forward in its wearable AI strategy. These glasses are designed to blend the digital and physical worlds, using Google’s Gemini AI to provide real-time assistance such as answering queries, offering translations, and understanding what the user is seeing through built-in cameras.

The upcoming demonstrations are likely to highlight how the smart glasses can function as an always-on AI companion, enabling hands-free interactions and more natural, context-aware support throughout daily tasks. With Android XR as the underlying platform, Google aims to make these glasses a key part of its broader vision for immersive, AI-driven computing.

Google is collaborating with several major partners, including Dell, HP, Lenovo, Asus, Qualcomm, and Intel, to develop the GoogleBook PC across a wide range of configurations. This approach is aimed at ensuring strong performance options and flexible hardware choices for different types of users.

By working with multiple manufacturers and chipmakers, Google plans to make the GoogleBook PC accessible across various price segments. This strategy is designed to meet the needs of users with different budgets while offering a consistent and optimized computing experience.

The US tech major is expected to showcase improvements in Google Beam, its AI-first 3D video communication platform. The system is powered by Google’s most advanced state-of-the-art AI volumetric video model, enabling highly realistic and immersive video calling experiences.

These enhancements are designed to make video calls feel more natural and lifelike, with improved depth, clarity, and presence. By leveraging advanced AI processing, Google Beam aims to transform standard video communication into a more engaging and interactive experience.

The company is expected to introduce a new version of Project Astra, an advanced AI system designed to better understand complex queries. It is built to handle more natural, long-form conversations with users, allowing it to interpret intent more accurately and respond more contextually and helpfully.

With these upgrades, Project Astra is aimed at completing multiple tasks more efficiently through a single, continuous interaction. This makes it easier for device owners to manage everyday actions without repeatedly issuing separate commands, significantly improving speed and user experience.

Google CEO Sundar Pichai and other senior executives are set to deliver keynote presentations, followed by developer sessions, live demos, and networking events. The conference will continue into day two on May 20 with additional talks and activities for attendees.

An Android 17 update is important because it represents the next step in improving how smartphones perform, look, and secure user data. Major Android releases typically bring better system efficiency, smoother performance, and new features that enhance everyday use, such as improved battery management, smarter AI integration, and upgraded privacy controls.

It also ensures apps remain compatible with the latest tools and security standards, helping protect users from emerging threats. For device manufacturers and developers, Android 17 sets new guidelines and capabilities that shape how apps and features are built for the next generation of smartphones.

Google I/O is important because it sets the direction for Google’s future products and technologies. It is where the company unveils major updates across Android, Google Search, AI, cloud services, and developer tools. For developers, it provides early access to new APIs, platforms, and frameworks that shape how apps are built and experienced.

For the broader tech industry, it signals emerging trends—especially in artificial intelligence and software innovation—that influence competitors and startups alike. Overall, Google I/O is a key event that showcases how Google plans to evolve its ecosystem and impact everyday digital experiences.

Google’s “Flash” Gemini models are built for speed and wide use across Android devices. This suggests future versions may focus more on fast response times than maximum performance, with even quicker models expected to be announced soon.

As in past editions, the event will include several focused sessions highlighting updates across key products and platforms, including Google Chrome, Android, and Google Cloud.

Chrome is expected to receive deeper integration with Gemini, bringing a range of built-in AI tools designed to make browsing more efficient. These features may include the ability to quickly summarise web pages, automatically fill out online forms, and handle routine browsing tasks with less manual effort, streamlining the overall user experience.

Google I/O 2026 LIVE updates: Google is also expected to showcase advancements in Android XR, including new Gemini-powered smart glasses and improved augmented and mixed reality experiences.

Google I/O 2026 LIVE updates: Android 17 is expected to bring smoother performance, better multitasking features, and more advanced smart tools integrated directly into the operating system.