Google's Gemini 3 is live after months of hype. Here's what it can do.

Core Insights - Google's Gemini 3 represents a significant advancement in AI capabilities, focusing on improved reasoning, creativity, and multimodal integration of text, images, and video [1][2][3] Group 1: Product Features - Gemini 3 enhances visual learning and reasoning, allowing for better explanation and visualization of ideas by understanding nuances across different media [3] - The model is designed to perform better searches by breaking down queries into more components, enabling the creation of visualizations and interactive graphics directly in search results [5] - Gemini 3 is claimed to be Google's most factual model to date, with a notable score of 37.5% on Humanity's Last Exam, indicating high reliability in solving math and science problems [6][7] Group 2: Availability and Access - Gemini 3 will be available immediately in Google Search for users who subscribe to the Pro or Ultra tiers, with plans to expand access to all users soon [4] - The introduction of Gemini Agent aims to facilitate multi-step tasks within the Gemini framework, enhancing user interaction with Google applications [8][9] Group 3: Unique Capabilities - The model features "latent capabilities," allowing it to create generative interfaces and interactive websites based on user queries [10] - A new platform called Antigravity is being launched, which leverages Gemini 3's coding abilities to automate tasks, such as building applications with minimal user input [11][12]