Google is all set to host its annual developer conference, Google I/O 2025, on May 20 and 21. Similar to last year, we can expect a wave of AI-focused announcements, including updates on Gemini, AI agents, Android XR, and more. In this article, we’ve rounded up the most anticipated features and announcements expected at this year’s I/O. Let’s dive in.
Gemini AI
At Google I/O 2025, the company is expected to unveil two new subscription tiers: Gemini Pro and Gemini Ultra. As reported by TestingCatalog, code strings referencing these plans have surfaced, suggesting they could offer expanded usage limits for features like video generation and more. Currently, the only paid tier available is Gemini Advanced, priced at $20 per month.
In addition, rumors suggest that Google might launch a new, more powerful Gemini Ultra model. The company hasn’t released an Ultra version since Gemini 1.0 Ultra. At present, Gemini 2.5 Pro stands as Google’s most advanced model, accessible to both Gemini Advanced subscribers and free users though the latter face usage limitations due to rate limits.
At Google I/O 2025, we’re likely to see the debut of three new AI models: Veo 3, Imagen 4, and Imagen 4 Ultra. Veo 3 will build upon the success of Veo 2, currently considered one of the best AI video generation models available. Meanwhile, Imagen 4 and its more advanced counterpart, Imagen 4 Ultra, will follow Imagen 3, Google’s cutting-edge image generation model.
Google’s AI Agent
At Google I/O 2025, the search giant is expected to finally release an AI agent for consumers. Back in December 2024, Google showcased Project Mariner, an assistant capable of performing actions within the Chrome browser much like OpenAI’s Operator AI agent. At the time, the company noted that the prototype was slow and not consistently accurate. Now, five months later, Project Mariner may be ready for a public launch.
In fact, some X users have discovered references to a feature called “Computer Use,” which enables Gemini to access the browser, execute code, and interact with the file system. This capability is reportedly geared toward enterprise users, allowing AI agents to operate within virtual desktop environments. With Google repeatedly stating that 2025 will mark the beginning of the agentic AI era, it’s highly likely we’ll see the debut of new and more advanced AI agents at I/O 2025.
Other Improvements Coming to Gemini
While Google already offers a personalized experience in Gemini by leveraging your search history, it may soon take things a step further at I/O 2025 by introducing memory and custom instructions. Early leaks suggest Gemini will receive a feature called “Personal Context,” which functions similarly to ChatGPT’s custom instructions.
With this, users will be able to specify their likes, dislikes, and provide personal details, allowing Gemini to better understand their preferences. Additionally, Gemini is expected to gain memory capabilities, enabling it to retain key details from conversations and build a long-term understanding of the user over time.
Next, Google might introduce the ability for users to upload files and images directly into the Deep Research AI agent on Gemini. Developers could also get the option to upload entire folders of their code repositories or integrate with GitHub to seamlessly bring large codebases into Gemini. Additionally, Google is reportedly working on a feature that will let users share Gemini Gems a concept similar to custom GPTs on ChatGPT.
Meanwhile, Google is also expected to unveil Gemini integration within the Chrome browser. Recent leaks suggest that Gemini will be accessible from the Windows system tray and can be launched using a dedicated keyboard shortcut. Additionally, the NotebookLM app will be available for installation on smartphones starting May 20.
Android XR
At I/O 2025, Google is expected to showcase its Android XR mixed-reality OS, aiming to compete with Apple Vision Pro and Meta Quest headsets. Samsung has already revealed that its ‘Project Moohan’ Android XR headset will launch this year. Notably, Google may also announce its own smart glasses, powered by the Android XR platform and Gemini AI.
Google’s smart glasses feature a sleek, lightweight design that looks like regular eyewear. They offer deep integration with Gemini AI, providing a seamless conversational experience. Equipped with advanced vision capabilities, the glasses can perform contextual tasks on the go, using Gemini AI to deliver an enhanced, immersive user experience.
Finally, Google unveiled its bold new ‘Material 3 Expressive’ design language with Android 16 just a week before Google I/O 2025. This marks the first time Google has hosted a separate Android event ahead of I/O. Additionally, Google confirmed that Gemini will be integrated into WearOS, Android Auto, Google TV, and Android XR platforms.
