OpenAI Unveils Operator, an AI Agent That Browses the Web Autonomously
OpenAI introduces Operator, a new AI agent that independently navigates web browsers to complete tasks like booking travel and shopping. The tool is initially available to U.S. ChatGPT Pro subscribers before expanding to other plans and regions.
OpenAI launches Operator, a new general-purpose AI agent that takes control of a web browser to independently perform tasks on behalf of users. The tool is currently available as a research preview for U.S. subscribers on ChatGPT's $200 Pro plan, with plans to expand to Plus, Team, and Enterprise tiers in the future. OpenAI CEO Sam Altman notes that international availability is coming soon, though European users face a longer wait due to regional regulations.
The Operator interface allows users to select from various task categories such as shopping, dining, delivery, and travel. When activated, a small window pops up to display a dedicated browser where the AI completes the requested chores while explaining its actions in real time. This dedicated browser setup ensures that users maintain full control of their own screens while the agent works in the background.
Powered by a new Computer-Using Agent (CUA) model, Operator combines the vision capabilities of GPT-4o with advanced reasoning to interact with website front-ends just like a human. Instead of relying on developer APIs, the AI navigates menus, clicks buttons, and fills out forms directly. OpenAI is actively partnering with companies like DoorDash, eBay, and Uber to ensure compliance with terms of service, and the system always asks for user confirmation before finalizing actions like submitting orders or sending emails.