Our most anticipated launch of the year is here.

• Gemini 3, our most intelligent model

• Generative interfaces, for perfectly designed responses

• Gemini Agent, made to complete complex tasks on your behalf

See how Gemini 3 can help you learn, build & plan anything 🧵

An AI powerhouse. Gemini 3 sets a new bar for AI model performance, so you can create, code, and build without missing a beat.

• Multimodal understanding, state-of-the-art reasoning and 1 million-token context window for synthesizing and analyzing videos, images, text, and https://t.co/HA4yyWrIO6

A new look. With a redesigned look and better response formatting, it’s easier to start chats, find what you’ve created and learn in a visual way.

• Information is more concise from Gemini 3, our most factual model.
• Two new experiments, visual layout and dynamic view, deliver https://t.co/29w8QHNJus

A Gemini Agent. With Gemini Agent, an experimental feature, Gemini can handle multi-step tasks for you—all under your guidance.

• Book a trip, organize your inbox or make appointments.
• Ask it to compare prices, reach out to vendors or schedule availability.
• Keep the control: https://t.co/zBIZJe09Vz

Free for students. U.S. students get Gemini’s Pro Plan free for an entire year, which includes access to Gemini 3, unlimited image uploads, Notebook LM and more. Sign up at https://t.co/NKA7mLpkdP

Gemini 3 is rolling out globally starting today. Try it at https://t.co/awhPeHZaAO, and select “Thinking” from the model picker.

Then let us know which features you’re most excited about in the replies.