Google I/O 2025 is the biggest conference for developers on the technological field. It takes place Tuesday and Wednesday at th Shoreline Amphitheatre in Mountain View and introduces the latest updates of Google world and announcements.
The event includes product announcements for Chrome, Google Search, YouTube, Android Studio, Gemini and all kinds of services related to the Mountain View enterprise. In this article we explore the latest news and rumors about the content for this year edition.
Google announcements for the I/O 2025
Android already had a separate event for announcements strictly related to the mobile operating system. The Android Show introduced new ways to find your lost phone, items, device-level features for the Advanced Protection program as well as security tools against scams and theft. Now it’s time for some general announcements in the Google I/O 2025 and we enlist the most expected ones.
Gemini Ultra
Only available in the U.S. for now, Gemini Ultra delivers the highest access to Google’s AI-powered assistant, apps and services. It costs 249,99 dollars month and it includes Google’s Veo 3 video generator. This is the latest Flow video editing app and it also works with the Gemini 2.5 Deep Think Pro AI capability mode that will be arriving soon.
AI Ultra rises the limits for Google’s NotebookLM platform as well as Whisk, the company’s image remixing app. Those who subscribe to AI Ultra can also access Gemini’s chatbot in Chrome. Other features include the Project Mariner tech, YouTube Premium and 30 TB of storage across the Google Drive, Photos and Gmail network.
Deep Think in Gemini 2.5 Pro and Google announcements
Deep Think is a reasoning mode for Google’s Gemini 2.5 Pro model with enhanced skills. It considers multiple answers to questions before responding and boosting the performance in several benchmark tests.
There are no details about how Deep Think works but experts have noticed similarities with OpenAI’s 1-pro and the upcoming o3-pro models. It will probably make use of an engine to search and synthesize the best possible solution for a given inquiry.
Trusted testers of Google can use Deep Think through Gemini API for now. The company is making some additional safety evaluations before launching Deep Think widely.
Veo 3 video-generating AI model experience
The Veo 3 model for video generation is one of the most interesting projects on the Google I/O 2025 roster. It can generate sound effects, background noises and dialogue for your video scenes. Veo 3 improves upon its predecessor Veo 2 in footage quality according to Google spokesmen.
The new feature will be available from Tuesday on through Gemini chatbot app for those who subscribe to the AI Ultra plan. Once you have it you only need to prompt it with a text or image and start creating.
Imagen 4 AI image generator
Google Imagen 4 is faster than Imagen 3 and it will get even faster in the near future. The release of a new Imagen 4 variant is imminent and it will boost speed 10x quicker than Imagen 3.
You can use this feature to render fine details like water droplets, fabrics and animal fur according to Google. Use Imagen 4 to create abstract style or photorealistic images in a wide range of aspect ratios and up to 2K resolution. It’s one of the tools to improve Flow, Google’s video tool pointed to filmmaking alternatives.
Gemini app updates
With more than 400 million monthly active users, Gemini apps are the star of the show for Google. Gemini Live’s camera and screen-sharing capabilities will arrive this week for iOS and Android users alike. It’s a feature powered by Project Astra to let people experience near-real-time verbal conversations with Gemini while they stream video from the smartphone’s camera or screen to the AI model.
The integration of Gemini Live will become deeper with other apps in the coming weeks. It will be able to offer directions from Google Maps, create events in Calendar or make a to-do list in Tasks. By updating Deep Research Gemini’s AI agent the user will also be able to upload private PDFs and images in just a few seconds.
Stitch Google I/O 2025 announcements
The AI-powered tool Stitch helps users to design web and mobile app front ends. It generates the necessary UI elements and code automatically. You can prompt Stitch to create app UIs using a few words description or an image. It provides HTML and CSS markup for each of its designs.
In comparison with other vibe coding products, Stitch is still limited. But there’s an interesting variety of customization options. The announcements from Google indicate that the AI agent Jules is also ready. It’s an AI solution to fix bugs in code and a very useful tool to understand complex code, create pull requests on GitHub and handle backlog items and programming tasks easily.
Project Mariner
This is one of Google’s experimental AI agents. It browses and uses websites to perform a dozen tasks at a time. You can purchase tickets for a football game or buy your groceries without visiting third-party websites. The user can just chat with Google’s Project Mariner and it will visit your websites and perform the actions you indicate.
Project Astra
Low latency, multimedia AI experience from Google. Project Astra powers an array of new experiences for Search, Gemini AI app and other products from third-party developers. It’s a tool to showcase nearly realtime, multimodal AI capabilities. Google says it’s building the Projet Astra glasses with other partners such as Samsung and Warby Parker. There’s not a release date yet.
AI Mode Google announcements
AI Mode is an experimental Google Search feature to perform complex multi-part questions using an AI interface. The AI Mode will make use of complex data in sports and finance queries and offer try it on options for apparel. Using Search Live you will be able to ask questions based on what your phone’s camera is seeing in real time. The Gmail platform will be the first to support personalized context indications.
Beam 3D teleconferencing
Starline it’s now called Beam. It employs a combination of hardware and software, including a six-camera array and custom light field display to converse with others. It can simulate being in the same meeting room with incredible results. The AI model converts video from cameras pointing from different angles and 3D rendering features.
Using Google’s Beam you can boast “near-perfect” millimeter-level head tracking as well as 60fps video streaming. In combination with Google Meet, Beam is an AI-powered real-time speech translation features preserving the original speaker’s voice, tone and expressions. The Google Meet platform will also receive real-time speech translation sooner.
More AI updates for technological fans
Google will launch Gemini in Chrome and give access to the newest AI browsing tool for free. It’s a tool that helps understating the context of a page and getting the tasks done. The Gemma 3n model can run “semoothly” on phones, laptops and tablets and it’s available from the preview section. It will support audio, text, images and videos.
Google also announced a ton of AI Workspace features arriving soon to Gmal, Google Dos and Google Vids. The most notable solution for Gmail is the personalized smart replies and a new inbox-cleaning feature. On the other hand, Vids gets new ways for content creation and edition.
Using the Video Overviews features that will arrive to NotebookLM and the SynthID Detect you can verify portals that use Google’s SynthID watermarking technology. It’s a way to indentify easily when the content is AI-generated. Lyria Real Time is the AI Model that powers the experimental music production app that you can download via API.
Wear OS 6
The Wear OS 6 program unifies font to tiles for a cleaner app look. For Pixel Watches you will get dynamic theming in order to synch app colors with watch faces. The main promise for the new design preference is that developers may build better customizations in apps to obtain seamless transitions. Google announcements include a design guideline and the Figma design files.
Google Play updates
The Play Store for Android is going to change. There will be new fresh tools t handle subscription, topic pages for the user to dive into specific interests. You can use audio samples and give folks a sneak peak into app content and the new checkout experience.
The topic browse pages for movies and shows will help connecting users to apps and use shows and movies alike. Plus, developers are getting dedicated pages for testing and releases. The tools to keep an eye and improve the app rollouts. Those developers that use Google can also now halt live app releases in case a critical problem pops up.
The subscription management tools are getting an interesting upgrade with multi-product checkout. Devs will be able to offer subscription add-ons as well as the main subscription method. All under the same payment.
Android Studio
The latest update for Google announcements includes agentic AI Journeys and the Agent Mode to handle more-intricate development issues. The new AI capabilities for Android Studio include a deep crash insights report. It can also recives the enhanced crash insights to make the best use of Gemini available.