Google updates the Gemini app with real-time AI video, Deep Research, and more

Google has announced several updates to the AI ​​chatbot application During Google I / O 2025More deep integration with more widely available multimodal AI features, updated AI models and Google product set.

Starting on Tuesday, Google is broadcast live camera and screen-sharing capabilities to all users on iOS and Android. The feature equipped with the Astra project is a video flow from the smartphone camera or screens simultaneously from the cell of the smartphone from the camera or screen to the AI.

For example, while walking a new city, users can point their phones in a building and have a response from the delay and delay, behind the Gemini.

In the coming weeks, Google says that Gemini Live will begin to integrate deeper with other applications. The company says that Gemini Live will soon offer directions from Google maps, it will be possible to operate on Google Calendar and Google Tasks.

Updates for Google’s Gemini, the company’s Openai is part of the efforts to compete with Chatgpt, Apple’s Siri and other digital assistants. The rise of AI Chatbots gave users a new way to interact with the Internet and their devices. This shows pressure on several large technologies, including Google Search and Google Assistant. Google, I / O Gemini has now announced in the period of 2025 now has a 400 million-month active user and the company hopes to grow the user base with these updates.

Google presented two new AI subscribers: Google Ai Pro, a rebrand for advanced plan for $ 20 twins per month and Google has ultraA $ 250 plan competing with Chatgpt Pro. Ultra Plan allows users to exclude access to very high rate limitations, new AI models and exclusive access to certain features.

Pro and ULTRA subscribers chosen as the language of the US chrome also announced the Gemini in Chrome browser, and declared Google Tuesday. Integration aims to ensure that users will answer questions about the generalizations of twin information or visible on their screens.

Google updates users to download private PDFs and images, Google, deep research, AI agent of the twins. Deep research will bring together these special PDFs with public data to create more individual reports. The company says that users will be able to quickly integrate into drivers and Gmail deep research.

Gemini’s free users, an updated AI image model that Google said better text outputs, Image 4 obtains. Subscribers to the new $ 250 AI Ultra Plan of the Company You can get sounds that match videos in Google’s latest AI video model, VEO 3, videos of video scenes.

Google also updates the default model to be a 25 Flash of Gemini, which will provide higher quality answers with a higher delay in the company.

To increase the number of students using AI chatbots, Google says Gemini will now be able to create individual quizzes on areas where users have difficulty. When users respond to questions wrong, the twins will help you create additional quizzes and action plans to strengthen these areas.