Google I/O 2025 Unveils Gemini 2.5, Project Astra, and Android XR
Google introduces its most advanced AI model yet, a real-time universal assistant, and a new extended reality platform at its annual developer conference.
Google kicks off its annual developer conference by unveiling Gemini 2.5, an upgraded AI model featuring a "Deep Thinking" capability for complex problem-solving. This new model seamlessly handles text, image, and audio inputs while powering updated features across Google Search, Workspace, and Android. Alongside the model release, Google introduces an AI Mode for Search that automatically generates comprehensive summaries for complex user queries.
Project Astra emerges as Google's vision for a universal AI assistant that processes visual and auditory inputs in real time. Demonstrated through an unedited live video, the assistant identifies objects, explains code, and recognizes surrounding locations. Google plans to bring these real-time multimodal capabilities to Pixel devices later this year through the Gemini Live voice assistant.
The tech giant also officially launches its Android XR platform, a dedicated operating system for extended reality devices developed in partnership with Samsung and Qualcomm. This initiative includes Android-powered smart glasses with built-in Gemini AI and a Samsung XR headset codenamed Project Moohan, both promising real-time translation and AI-driven AR overlays. Additionally, Google showcases Veo and Imagen 4, two new creative AI models designed for generating high-quality cinematic video and ultra-realistic images from text prompts.