Google supercharges Gemini with new AI options at I/O 2025: 13 Updates you shouldn’t miss | Mint-OxBig News Network

Google has rolled out a set of main updates to its Gemini app, aiming to widen its enchantment and supply customers extra methods to work together with AI. Announced on the annual Google I/O 2025 occasion on Tuesday, the updates embrace instruments for visible assist, media era, and analysis — now obtainable on each Android and iOS. Here is each AI replace introduced by the American tech large:

Gemini Live

Gemini Live, which lets customers share their digital camera feed or display throughout conversations, is now obtainable without spending a dime. The characteristic is designed to assist customers present what they imply as an alternative of typing out questions. According to Google, conversations utilizing Gemini Live are typically longer than text-only chats, which the corporate attributes to the extra interactive format.

Google plans to combine the instrument extra deeply into its wider ecosystem within the coming weeks. Users will be capable to hyperlink Gemini Live with apps like Maps, Calendar, Tasks, and Keep. For instance, asking about restaurant choices could hyperlink on to Google Maps, or a gaggle chat might result in an occasion being added to Calendar.

Imagen 4

Gemini now consists of Imagen 4, a brand new picture era mannequin that helps improved visible element and higher textual content rendering inside pictures. Users can create graphics and visuals for numerous makes use of, together with displays and social media posts.

Veo 3

For video creation, Veo 3 is being launched. It helps text-to-video era and also can add ambient sounds and fundamental character dialogue. Veo 3 is presently obtainable solely toGoogle AI Ultra subscribers within the U.S., limiting entry for worldwide customers or these on the free plan.

Deep Research

The Deep Research characteristic now permits customers to add private information — equivalent to PDFs or pictures — to be included in AI-generated reviews. The aim is to offer extra personalised and context-rich outcomes by combining non-public and public knowledge sources. Google has introduced plans to develop this performance to incorporate content material from Google Drive and Gmail within the close to future.

Project Astra

Project Astra showcases the real-time capabilities of Google’s Gemini fashions, with the preliminary options now built-in into Gemini Live. This upgraded model can actively use a tool’s digital camera to interpret on-screen content material in real-time. Among the newest enhancements are a extra pure, expressive voice powered by native audio era, enhanced reminiscence performance, and superior pc management options.

During the keynote at Google I/O 2025, a reside demonstration highlighted Gemini Live’s means to work together fluidly with customers—responding with expressive speech, dealing with interruptions seamlessly, and persevering with conversations with out shedding context. It additionally showcased multitasking skills equivalent to making enterprise calls, scrolling by way of paperwork, and looking the online, all in real-time.

Google Flow

Google Flow is an AI-powered filmmaking instrument designed for creatives to effortlessly generate cinematic movies. It combines Google’s superior fashions — Veo (video era), Imagen (picture era), and Gemini (pure language understanding) — to assist customers flip on a regular basis language prompts into high-quality visible scenes. Flow permits constant character and scene creation, permitting seamless integration throughout a number of clips. It’s constructed to make storytelling sooner, intuitive, and visually beautiful.

Agent Mode

At the Google I/O 2025 occasion, Pichai unveiled a brand new characteristic known as Agent Mode for the Gemini app. This upcoming experimental instrument, initially obtainable to subscribers, is designed to deal with advanced duties and planning on the consumer’s behalf. With Agent Mode, Gemini strikes past easy responses to tackle extra autonomous actions—organising, scheduling, and executing multi-step duties. Google additionally introduced that these agentic AI capabilities will prolong to Chrome, Search, and the Gemini platform, marking a major step towards AI that may proactively handle duties relatively than simply react to prompts.

Google Jules

Jules is an autonomous, agentic coding assistant that works instantly along with your codebase. Unlike conventional code-completion instruments, Jules clones your repository right into a safe Google Cloud VM, understands your mission context, and independently handles duties like writing checks, fixing bugs, constructing options, and extra. It works asynchronously, so you possibly can focus elsewhere whereas it completes duties and returns with an in depth plan, reasoning, and code adjustments. Jules is now in public beta and prioritises privateness, retaining your code safe and remoted.

AI Mode in Search

Google is rolling out a brand new characteristic known as AI Mode, aimed toward individuals who need a extra superior and interactive search expertise. First examined in Labs, AI Mode is now obtainable to everybody within the U.S., with a wider international rollout anticipated later. A brand new tab for AI Mode will quickly seem within the Google app and on desktop.

AI Mode makes use of one thing known as a “query fan-out” system. This means it breaks down your query into smaller elements and runs many searches directly, serving to it dig deeper and return extra helpful and detailed solutions from throughout the web. It additionally makes use of Gemini 2.5, Google’s most superior AI mannequin but.

With AI Mode, customers can ask follow-up questions, get interactive hyperlinks, and even use pictures or reside video to look in real-time. It is not only about answering questions anymore, Google desires to assist folksdo issues, from reserving tickets to evaluating knowledge.

Google Meet speech translation in real-time utilizing AI

Google has launched a groundbreaking AI-powered speech translation characteristic in Google Meet, enabling real-time audio-to-audio translation throughout calls. Built on DeepThoughts’s superior AudioLM expertise and built-in with the Gemini AI mannequin, this method interprets spoken language right into a listener’s most popular language—whereas preserving the speaker’s authentic voice, tone, and emotional expression.

Unlike conventional caption-based translation, this characteristic instantly transforms speech, delivering natural-sounding audio in actual time. Users hear the translated voice with delicate overlays of the unique, enhancing readability and sustaining conversational context. Though there’s a slight delay for processing, the expertise intently mimics having a reside interpreter on the decision.

Google Beam

Google Beam is a brand new 3D video communication platform that transforms common 2D video calls into immersive 3D experiences. Announced at Google I/O 2025, it makes use of a number of cameras and AI to create lifelike, real-time 3D visuals with depth and eye contact. Powered by Google Cloud and designed for enterprise use, Beam additionally helps exact head monitoring and is anticipated to characteristic real-time speech translation. It will roll out on HP gadgets later this 12 months.

Gemma 3n

Gemma 3n is Google’s first open AI mannequin designed particularly for on-device use, bringing quick, multimodal intelligence to telephones, tablets, and laptops. Built on a brand new structure developed with companions like Qualcomm, MediaTek, and Samsung, it powers real-time, non-public AI experiences with out counting on the cloud. Gemma 3n additionally types the inspiration for the subsequent era of Gemini Nano, enabling builders to discover superior AI instantly on on a regular basis gadgets.

Try-on

The “Try On” characteristic in Google Search permits customers to see how garments like shirts, clothes, pants, and skirts would look on them by importing a full-length picture. Available by way of Search Labs within the U.S., the instrument makes use of AI to generate a visible of the outfit on the consumer. It additionally lets customers save or share the photographs for suggestions earlier than making a purchase order.