To carry out tasks on the web, Operator uses a new model called Computer-Using Agent (CUA), which combines GPT-4o's vision skills with more advanced reasoning. That combination lets it interact with the menus, controls, and text fields on a web page without the need for a custom API.

OpenAI has released a preview of its first AI agent, dubbed Operator, that can automate “repetitive browser tasks such as filling out forms, ordering groceries, and even creating memes." At launch, Operator is available to those with a $200-per-month ChatGPT Pro subscription in the US,

OpenAI is delivering on its promise of making 2025 the year of agentic AI. Last week, the company launched Tasks for ChatGPT, which lets users automate future prompts like sending

In the meantime, the user can either watch the Operator work on the web or do something completely different. In some cases, the Operator will wait for the user — for example, if it needs to log in somewhere or confirm that an email should be sent.