Google I/O 2025, the largest developer conference in this year, will be held on Tuesday and Wednesday at the Shoreline runway at Mountain View. On the ground we offer you the latest updates from the event.
I/O displays products from all Google wallets. We have a lot of news related to Android, Chrome, Google Search and YouTube, and of course-Chatbot on behalf of artificial intelligence, Gemini.
Google hosted a separate event intended for Android updates: Android offer. The company has announced new ways to find lost Android phones and other elements, additional features at the device level of the advanced protection program, safety tools for protection from fraud and theft, and a new design language called Mateive 3 Expressive.
Here are all the things announced in Google I/O 2025.
Gemini Ultra
Gemini Ultra (only in the United States currently) provides “the highest level of access” to Google applications and services that work with AI, according to Google. Its price is $ 249.99 per month and includes the Google Veo 3 video generator, the company’s new video editing app, and a strong capacity for artificial intelligence called GEMINI 2.5 Pro Deep Think Mode, which has not yet been launched.
AI Ultra comes with higher borders on the Google and WestKLM Notebooklm platform, which is the company to redesign the company’s photo designing. AI Ultra subscribers can also access the Google Chatbot Gemini in Chrome; Some “Agentic” tools are supported by the company’s Mariner Tech project; Youtube Premium. And 30 TB of storage via Google Drive, Google Photos and Gmail.
Deep thinking in Gemini 2.5 Pro
Deep Think is the “improved” thinking mode of the Gmini 2.5 Pro pioneering model from Google. The model is allowed to consider multiple answers to the questions before answering, and enhancing its performance on certain criteria.
Google did not inspect details about Work Think Works, but it may be similar to OPNAI OPNAI models and upcoming O3-PRO models, which are likely to use a driver to search for the best solution for a specific problem.
Deep thinking is available to “trusted laboratories” through the Gemini Application Programming interface. Google said it takes an additional time to make safety assessments before a widespread deep thinking.
Veo 3 AI in VEO 3
Google claims that Veo 3 can create sound effects, background noise, and even a dialogue to accompany the videos that create. Google says VEO 3 also improves its predecessor, VEO 2, in terms of the quality of the shots that can create it.
VEO 3 is available on Tuesday in the Goeni Chatbot app for the AI Ultra plan, which is valued at $ 249.99 in Google, where it can be demanded for text or image.
Image 4 Amnesty International is a photo generator
According to Google, Imagen 4 is fast – faster than imagen 3. It will become faster. In the near future, Google plans for a variable version of IMAGEN 4 that reaches 10x faster than Imagen 3.
Imagen 4 is able to provide “fine details” such as fabrics, water drops, and animal fur, according to Google. It can deal with both realistic and abstract patterns, and create pictures in a range of display rates to height and accuracy of 2K.
Both VEO 3 and Imagen 4 will be used for energy flow, which is a video tool on behalf of artificial intelligence directed towards filmmaking.
Gemini application updates
Google has announced that Gemini applications have more than 400 active users per month.
The Gemini Live camera and screen sharing capabilities this week will be shown to all users on iOS and Android. This feature, with the Project ASTRA, allows people to have oral conversations close to time with GEMINI, while they also broadcast the video from a camera or their smartphone screen to the artificial intelligence model.
Google says that Gemini Live will also start integrating deeper with its other applications in the coming weeks: You will soon be able to provide trends of Google Maps, create events in Google Calendar, and to provide lists of tasks with Google.
Google says she is updating deep research, AI’s agent in Gemini who generates comprehensive research reports, by allowing users to download PDF files and their photos.
stitch
STITCH is an AI energy tool to help people design Web and App Front Front by creating the necessary user interface elements and symbol. STITCH can be required to create the App application with a few words or even a picture, providing HTML and CSS for the designs it generates.
Stitch is slightly more limited as it can do compared to some other coding products, but there is a good amount of customization options.
Google has also expanded the scope of access to JULES, AI’s agent aims to help developers fix errors in the code. The tool for developers help to understand complex software instructions, create withdrawals on GitHub, and deal with some of the accumulation and programming tasks.
Mariner Project
Project Mariner is a Google experimental AI agent that browses and uses websites. Google says she has greatly updated how Project Mariner works, allowing the agent to take nearly ten tasks at one time, and is now rotating it to users.
For example, the Project Mariner users can buy tickets to the baseball game or buy groceries via the Internet without visiting a website to an external entity. Only people can chat with a Google AI agent, visit web sites and take action for them.
Astra Project
The Google Low Access Time experience, the AI multimedia experience, the ASTRA project will run a set of new search experiences, the Gemini Ai application, and the products of the third party developers.
Project Astra was born of Google DeepMind as a way to display almost multimedia capabilities. The company says it is now building these ASTRA glasses with partners including Samsung and Warby Parker, but the company does not have a specific launch date yet.
Artificial intelligence mode
Google poses AI’s mode, which is the Google Experimental Search feature that allows people to ask multi -portable questions across the AI interface, for users in the United States this week.
Artificial intelligence mode will support the use of complex data in mathematical queries and financing, and will provide “try it on” clothes. Live search, which is asked later this summer, will allow you to ask questions based on what your phone camera sees in actual time.
Gmail is the first application to be supported in a personal context.
3D beams remote conferences
Beam, previously called Starline, uses a set of programs and devices, including a set of six cameras and a light screen, to allow the user to talk to someone as if in the meeting room itself. The artificial intelligence model turns a video of the cameras, which is placed in different angles and directed towards the user, into a 3D width.
Google beam is proud to track “almost perfect” millimeter and a video flowing 60 frames per second. When using it with Google Meet, Beam provides the ideological translation feature of Amnesty International that maintains the sound, tone and expressions of the original speaker.
And talking about Google Meet, Google announced that Meet is to get the actual time translation.
More artificial intelligence updates
Google launches Gemini at Chrome, which will allow people to access a new AI browsing assistant quickly helps them understand the context of the page and get tasks.
GEMMA 3N is a model designed to run “smoothly” on phones, laptops and tablets. It is available in the inspection starting on Tuesday; He can handle sound, text, photos and videos, according to Google.
The company has also announced many features of the update intelligence space to Gmail, Google Docs and Google Vids. It is worth noting that Gmail gets custom smart responses and a new feature to clean the incoming mail, while Vids gets new ways to create and edit content.
Video overview of Notebooklm, and the company has launched Synthid detector, which is the verification portal that uses the Google SYNTHID Waterming technology to help identify the content generated from artificial intelligence. Lyria Realtime, the artificial intelligence model that operates the experimental music production application, is now available via the application programming interface.
Wearing OS 6
Wear OS 6 brings a unified line of tiles to look at the cleaner app, and the pixel watches receive dynamic clothes that coincide with watches.
The main promise of the new design reference platform is to allow developers to build a better customization in applications along with smooth transitions. The company launches Design guide For developers, along with Figma design files.
Google Play
Google reinforces the Play Store for Android developers with new tools to deal with subscriptions, and the topic pages so that users can dive into specific interests, audio samples to give people a peek on the content of the application, and the new exit experience to make the sale of additional functions more smooth.
“Browse the topic” pages for movies and shows (only for us now) connect users to applications associated with tons of shows and movies. In addition, developers get tests for testing and versions, and tools needed to monitor and improve the deportation of their applications. Devils who use Google can also stop the Live application versions if a critical problem appears.
Subscription management tools also get a upgrade with multi -product exit. Devs will be able to provide additional jobs to subscribe along with the main subscriptions, all under one payment.
Android Studio
Android Studio combines new artificial intelligence features, including “trips”, the “Agentic Ai” capacity that coincides with the Gemini 2.5 Pro model. The “agent” will be able to deal with more exciting development processes.
Android Studio will receive new potential of artificial intelligence, including the improved “Crash Insights” feature in App Application Quality Insights. This improvement from Gemini will analyze the application source code to determine the possible causes of accidents and suggest repairs.
