Gemini Ultra, Veo 3, and Imagen 4 were presented at Google I/O ...
Google has unveiled a range of new developments in artificial intelligence at this year's I/O conference, held in Mountain View, California on May 21-22, 2025. The conference showcased various AI-related innovations, including advancements in photo and video generation, updates to the Gemini family, new AI tools for UI and code creation, and more.
Google Beam and Subscription Plans
One of the highlights of the conference was the introduction of Google Beam, an AI-powered video communication platform that converts 2D video into immersive 3D experiences. Google Beam offers high-quality head tracking and operates in real-time at 60 frames per second. The initial rollout of Google Beam will be in partnership with HP.
In addition, Google announced a new subscription plan for its AI services, offering premium access to advanced models and features. This plan, initially available in the US at $249.99 per month, provides users with enhanced capabilities and higher limits.
Real-Time Language Translation with Google Meet
Google Meet now features real-time language translation for Google AI Pro and Ultra subscribers, starting with English and Spanish translations. The platform aims to preserve voice intonation and emotions during translations, providing a seamless communication experience.
Updates to Gemini Models
Google introduced new features to its Gemini 2.5 Pro and 2.5 Flash models, including built-in audio output for natural conversations, enhanced security features, and computer-assisted functionalities from Project Mariner. The Gemini 2.5 Pro model also received Deep Think, an experimental mode for advanced problem-solving capabilities.
Introducing Flow and Veo 3
At the conference, Google unveiled Flow, an AI-based cinematography tool that integrates various models such as Veo, Imagen, and Gemini to create compelling visual content. Flow is currently available to Google AI Pro and Ultra subscribers in the US.
Veo 3, the latest addition to Google's AI video creation family, offers advanced capabilities to generate not only video but also sound effects. This model is accessible to Ultra plan subscribers in the US and corporate users of Vertex AI.
Imagen 4 and Other Innovations
Imagen 4, the newest AI model for image generation, specializes in photorealistic rendering and detailed visual representation. It is compatible with Gemini, Whisk, Vertex AI, and various Workspace tools.
Furthermore, Google showcased Lyria 2 for music creators, introduced Project Astra for universal AI assistance, and unveiled new tools like Stitch for designing interfaces based on Gemini 2.5 Pro.
Project Mariner and Future Developments
Google's ongoing projects include Project Mariner, which enables AI agents to perform multiple tasks simultaneously, enhancing user experiences. The company is also working on Project Astra glasses and expanding access to AI tools like Jules for developers.
As Google continues to push the boundaries of AI technology, these new developments promise exciting possibilities for the future of artificial intelligence and human-machine interactions.
