Google CEO Sundar Pichai shares 5 major Gemini 3 upgrades you need to know about

Google CEO Sundar Pichai has highlighted the powerful capabilities of twin 3, the companion’s in style(p) and to the highest degree innovative AI model. In a post on X on Wednesday, he explained five new features that aim to make everyday tasks easier, faster, and more interactive for users. Here is a simple breakdown of what Gemini 3 can do, according to Pichai.

1. Gemini can turn almost anything into something new

Pichai says Gemini 3 can understand many types of inputs, such as photos, PDFs, rough sketches, and diagrams. You can show it as a doodle on a napkin, and it can turn it into a full website. A simple picture can become a board game, and a diagram can be turned into a lesson you can interact with.

2. Google’s AI now understands videos much better

The new model can watch long videos and break them down for you. Pichai says it can analyse sports videos, point out what you are doing wrong, and even suggest drills to help you improve. This is because Gemini 3 now has stronger visual and spatial reasoning.

3. How is Gemini 3 making Google Search more helpful?

According to the Google CEO, Gemini 3 is also improving the way Search works. Instead of just giving written answers, the model can create visual layouts and interactive tools. For example, if you ask about the three-body problem in physics, you may get a simulation that helps you understand how it works.

4. Gemini 3 gives results that look like a digital magazine

Search results can now appear in a more visual and engaging way. Gemini 3 can show photos, modules, and interactive sections that you can tap through. If you ask it to plan a three-day trip to Rome, you get a neat, scrollable itinerary tailored to your likes.

5. AI can take actions for you

Google is introducing Gemini Agent, a smart tool that helps with everyday chores. It can organise your email, draft replies, archive old messages, or help book local services. The assistant figures out the steps on its own and suggests useful actions. It will be available on the web for Google AI Ultra subscribers in the United States.

How powerful is Gemini 3?

Google explains in its blog post that Gemini was designed to understand many kinds of information at the same time, such as text, images, videos, audio and even computer code. The new Gemini 3 takes this ability even further. It can reason better, see and understand things more clearly, work in many languages, and handle very long inputs all at once.

For example, if you want to learn cooking passed down in your family, you can give Gemini 3 old handwritten recipes, even if they are in different languages. It can read them, translate them and turn them into a clean, easy to share family cookbook.

If you want to learn something new, you can share research papers, long video lectures or tutorials. Gemini 3 can turn all of that into useful tools like interactive flashcards, charts or visuals that make it easier to understand.

It can also watch a video of you playing pickleball, point out what you need to improve and create a full training plan to help you get better.

---

---