Brand Makers
Dil Ka Jod Hai, Tootega Nahin
Sundar Pichai unveils key upgrades in Google’s new Gemini 3 model
In a supporting blog post, Google noted that Gemini was built to process multiple types of information simultaneously—from text and images to video, audio and code.
In a supporting blog post, Google noted that Gemini was built to process multiple types of information simultaneously—from text and images to video, audio and code.
Google CEO Sundar Pichai has outlined five significant upgrades introduced in Gemini 3, the company’s most advanced AI model to date. In a post on X, he detailed the features designed to make daily tasks faster, more intuitive and highly interactive, while marking a substantial leap in multimodal AI capability.
Pichai stated that Gemini 3 can now turn almost any type of input into something entirely new. The model is capable of interpreting photos, PDFs, rough sketches and diagrams, and transforming them into full outputs such as websites, board games or interactive lessons. He emphasised that even a hand-drawn doodle on a napkin can become a fully functioning web page.
1/ Gemini 3 in action, a quick????
— Sundar Pichai (@sundarpichai) November 18, 2025
You can give Gemini 3 anything (images, pdfs, scribbles, etc) and it will create whatever you like: an image becomes a board game, a napkin sketch transformed into a full website, a diagram could turn into an interactive lesson. pic.twitter.com/xCry35mFbW
He added that Gemini 3 now understands videos with far greater accuracy. The model can watch lengthy footage, break it down into meaningful analysis and even offer personalised guidance. Pichai noted that it can review sports videos, identify errors and suggest drills for improvement, owing to its stronger visual and spatial reasoning abilities.
The Google CEO also explained how Gemini 3 is improving Google Search. Instead of providing only text-based answers, the model can generate visual layouts and interactive tools. As an example, he said a query related to the three-body problem in physics could produce a simulation that makes the concept easier to grasp.
Search results themselves have also been redesigned to appear more like a digital magazine. With Gemini 3, users receive visually rich responses featuring images, interactive modules and scrollable sections. Pichai highlighted that asking for a three-day itinerary in Rome would yield a neatly structured plan tailored to individual preferences.
The fifth upgrade centres on Gemini Agent, a new assistant that can take practical actions on a user’s behalf. Pichai said it can organise email inboxes, draft responses, archive old messages and help book local services. The tool determines the required steps on its own and suggests useful actions. Gemini Agent will be available on the web to Google AI Ultra subscribers in the United States.
In a supporting blog post, Google noted that Gemini was built to process multiple types of information simultaneously—from text and images to video, audio and code. Gemini 3 advances this significantly, with improved reasoning, sharper visual understanding, broader language capability and the ability to handle very long inputs at once.
SPOTLIGHT
"The raucous, almost deafening, cuss words from the heartland that Piyush Pandey used with gay abandon turned things upside down in the old world order."
Read MoreThe new face of the browser: Who’s building AI-first browsers, what they do and how they could upend advertising
From OpenAI’s ChatGPT-powered Atlas to Microsoft’s Copilot-enabled Edge, a new generation of AI-first browsers is transforming how people search, surf and interact online — and reshaping the future of digital advertising.