.png)
The Google I/O 2025 developer conference keynote on Tuesday was jam-packed. During the event, firm CEO Sundar Pichai and other executives unveiled a slew of new artificial intelligence (AI) improvements and capabilities. Some of them include additional features in the Gemini 2.5 series of AI models, changes to AI Mode in Search, expanded AI Overviews, the debut of the new 3D communication platform Google Beam, and a demonstration of the Android XR platform. In case you missed the live event, here's a concise recap of everything revealed.
Google Beam
Google's Project Starline is currently known as Google Beam, a 3D communications platform. It employs a series of six cameras to record a video feed of the user from various perspectives. Then, an AI system integrates these to convert the 2D stream into a 3D light field display. The business also employs head-tracking sensors to correctly record the user at 60 frames per second (fps).
Google is partnering with HP to launch the first Google Beam devices later this year. Only a restricted group of buyers will receive the initial devices. Additionally, Google Beam devices from original equipment manufacturers (OEMs) will be accessible via InfoComm 2025, which is scheduled for June.
Gemini 2.5 Upgrades
The Gemini 2.5 series is also receiving some new features. The 2.5 Pro model will now have a new Deep Think mode, dubbed an advanced reasoning mode. The functionality is presently being tested. Native Audio Output, which allows for expressive and human-like voice creation, is also being enabled to Gemini 2.5 models via the Live application programming interface (API).
Google is also improving the Gemini 2.5 Flash model with new features for reasoning, multimodality, coding, and lengthy context. The model will also be more economical to use. Developers utilizing the Gemini API will also receive thought summaries and budgets for the most recent models.
AI Mode in Search.
Another hot topic from the keynote session was the AI Mode in Search. Google intends to power the end-to-end AI search using a proprietary Gemini 2.5 model. The AI Mode will also receive a new Deep Search mode, a Live Search feature that allows the AI tool to access a device's camera, and a new agentic function that allows users to purchase event tickets and schedule appointments straight from the interface.
AI Mode in Search is also receiving some new shopping-related capabilities. Users will now be able to visually search for the product they desire, try on a broad range of items digitally by uploading a photo of themselves, and utilize AI agents to track product pricing and make purchases automatically. These features will be implemented later in the year.
AI Overviews: Expansion
During the speech, the Mountain View-based tech behemoth announced the extension of AI Overviews. The AI-powered search result snapshot function will now be accessible in more than 200 countries and 40 different languages. Arabic, Chinese, Malay, and Urdu will be added to the list of supported languages, which now includes English, Hindi, Indonesian, Japanese, Portuguese, and Spanish.
Gemini-powered Android XR
During the speech, Google also demonstrated the new Gemini-powered Android XR platform. It will serve as the operating system for Samsung's future Project Moohan smart eyewear. The firm also collaborates with other wearable partners.
These Android XR smart glasses will have a camera, microphone, speakers, and an over-the-glass display panel. Users will be able to communicate with Gemini hands-free, request image capture, operate their smartphone and other linked devices, and much more.
Imagen 4 and Veo 3 AI models
Google also announced Imagen 4, its next-generation image generation model, and Veo 3, its video generation model. Imagen 4 now has enhanced text rendering and contextual comprehension of word arrangement, as well as higher image quality and faster adherence.
Veo 3 introduces native audio creation capability, which implies that created movies will now include ambient noises, background music, and conversations. Both versions will be available to the public later this year.
In addition, the business is unveiling Flow, a new AI-powered filmmaking tool. It uses Imagen, Veo, and Gemini to create eight-second video segments. Multiple clips can be stitched together to form a larger scenario. The software supports both text and graphics as prompts.
Gemini on Google Chrome
Paid members will now be able to use the Gemini AI assistant in Google Chrome. A new Gemini button will allow visitors to summarize a webpage or ask questions about its content. It can also traverse webpages automatically depending on user input. The AI assistant can also handle many tabs at the same time.
The Stitch Tool
Google also introduced a new AI-powered tool for creating app interfaces based on text prompts and templates. The program, called Stitch, also accepts wireframes, preliminary drawings, and screenshots of other user interface (UI) designs. It is now accessible as an experiment through Google Labs.
Speech Translation for Google Meet
In addition, Google Meet will now have a new AI capability. The video conferencing platform will now feature real-time voice translation, allowing speakers of various native languages to chat with little latency. Currently, the function can translate English and Spanish. It is presently under beta testing with premium subscribers.
Google AI and AI Ultra plans
Finally, Google announced the Google AI and Google AI Ultra plans for its Gemini features. The former replaces the Google One AI Premium plan and will cost $19.99 (Rs. 1,950 in India) per month, while the Google AI Ultra plan will cost $249.99 (approximately Rs. 21,000) per month. The latter will receive all new services first, with greater rate limitations and 30TB of cloud storage.
