Google announced updates to the Gemini app, including Imagen 4 and Veo 3 access

This Tuesday at I/O 2025, Google didn't hold back with the AI-related announcements. In addition to sharing details about further updates to the Gemini 2.5 model lineup, the company also advertised several exciting new features coming to its Gemini AI assistant, like new subscription tiers and enhanced capabilities that signal the company's push toward more personalized and proactive AI experiences.

The headlining announcement is that Google is making Gemini Live with camera and screen sharing available free for all users on Android and iOS. This feature allows users to point their phone at objects for real-time visual assistance. According to Google, users with access to Gemini Live engage in conversations that are reportedly five times longer than text-based interactions.

To build on Gemini Live's current success, Google is introducing deeper integrations with its ecosystem, starting with Google Maps, Calendar, Tasks, and Keep. These integrations will enable Gemini to take action, such as setting aside a date in Google Calendar when users discuss details of upcoming plans.

Gemini is also getting model updates across the board: 2.5 Flash is the new default model for the Gemini assistant, and users can now leverage the Imagen 4 integration for high-quality image generation. Additionally, Google is rolling out its video generation model, Veo 3, for subscribers of its new Google AI Ultra subscription plan. Veo 3 boasts support for native sound effect generation, enabling users to generate immersive videos with dialogue, background noise, and sound effects from simple text prompts.

Additional updates include private file upload support for Deep Research (Google app connectors are upcoming), an improved Canvas experience allowing users to collaborate with Gemini in everything from infographics and quizzes to working code generation, Chrome integration for web browsing assistance, and interactive quizzes for students.

Google has also launched two new subscription plans: Google AI Pro at $19.99/month, aimed at replacing and expanding the Gemini Advanced subscription, and Google AI Ultra at $249.99/month for power users seeking the highest rate limits and early access to experimental features. Starting this April, college students from select countries can sign up for a free Gemini update through free access to the Google One AI Premium plan, valid until spring 2026.

Remarkably, Google AI Ultra users will have early access to new models and enjoy exclusive features, such as an experimental capability called Agent Mode. As the name reveals, Agent Mode is Google's take on an assistant that can take on actions on behalf of the user without requiring constant supervision. As Google describes it, Agent Mode "seamlessly combines advanced features like live web browsing, in-depth research and smart integrations with your Google apps, empowering it to manage complex, multi-step tasks from start to finish with minimal oversight from you."