OpenAI has launched ChatGPT Agent, a new AI-powered tool that enables ChatGPT to autonomously perform complex, multi-step tasks by controlling a virtual computer. The agent integrates capabilities from OpenAI's previous tools, including Operator's action-taking web browser, Deep Research's web synthesis, and ChatGPT's conversational intelligence. It can browse the internet, run code, fill out forms, create documents such as PowerPoint presentations and spreadsheets, shop online, and connect to applications like Gmail and GitHub. The ChatGPT Agent is designed to complete tasks from start to finish while maintaining context across activities, requiring user permission before executing real-world actions. Initially, the feature is available to Pro, Plus, and Team subscribers. OpenAI describes this as a new level of AI capability, combining research and operational functions into a unified system to enhance productivity and task automation.