Google I/O 2025: Gemini, Search, Smart Glasses

Google I/O is considered by many software developers to be one of the most important developer conferences. Google has grown beyond its original focus of Android, Chrome, Search, and Workspaces, and AI is considered the next area of advancement. Here's a summary of the major announcements from Google I/O 2025, covering Gemini updates and other services that may change how we communicate and collaborate.

Gemini Updates

Google released Gemini 2.5 Pro earlier than expected, after positive feedback from beta testers. Google also announced improvements to the Gemini 2.5 Pro and Flash models, including Deep Think support for Pro, enabling high-level research techniques. Gemini 2.5 Flash has improved reasoning, multimodality, coding, and response efficiency, using 20% to 30% fewer tokens, according to Google. Gemini 2.5 models will support audio-visual input and native audio out dialogue via a preview version in the Live API.

Project Mariner and AI Subscriptions

Google aims to build an AI using Gemini 2.5 Pro that is intelligent, aware of context, and capable of planning and acting across devices. Project Mariner, a browser-based AI, can handle tasks like booking flights, research, and shopping. The latest prototype will be available to Google AI Ultra subscribers in the US first.

Google is offering two AI subscriptions: Google AI Pro and Google AI Ultra. The Pro plan includes AI products with higher rate limits and special features. The Ultra plan includes Google's highest rate limits and early access to experimental AI products like Project Mariner, Veo 3, and Gemini 2.5 Pro with Deep Think mode. The Ultra subscription also includes early access to Agent Mode for desktop-level prompts.

Generative Media Models

Google introduced new versions of its generative media models, including Veo 3, Imagen 4, and Flow. Veo 3 now supports audio prompt generation. Flow, an AI filmmaking tool, allows creators to adjust camera angles, motions, cast, and location. Imagen 4 has improved accuracy and clarity, especially in details like textures and fur. It can also produce content in various aspect ratios and up to 2K resolution.

AI Mode in Search

AI Mode in Search will be widely available in the US, without requiring Labs sign-up, and will include new capabilities. Deep Search in AI Mode expands background queries to provide more thorough search responses. The Google Search experience is also gaining Project Astra's multimodal capabilities. Project Mariner's agentic functions will also be added to AI Mode. Google has expanded AI Overviews to over 200 countries and territories and more than 40 languages.

Workspace Updates

Google is adding AI features to Workspace services like Gmail, Meet, and Docs. Gmail is introducing personalized smart replies and an inbox cleanup feature. Gemini will suggest meeting booking windows in Gmail threads. Google Meet will offer near-real-time speech translation for Google AI Pro and Ultra subscribers. Docs users can establish source-grounded writing help.

Google Beam

Project Starline is now Google Beam. Google Beam uses AI to transform 2D videos into 3D experiences. Google is partnering with HP to create communication devices running Google Beam.

Android XR Glasses

Google is focusing on Android XR glasses that use cameras, microphones, and speakers to interpret visuals and assist with Gemini. With Android XR's Gemini capabilities, users can access directional systems, respond to texts, translate conversations, and capture photos via voice commands. Google is partnering with eyewear brands like Gentle Monster and Warby Parker to create stylish smart glasses, expected later this year.