OpenAI launches Operator, an AI agent to help you perform tasks autonomously
It can automate tasks like flight books, making hotel reservations, shopping online, and so much more.
While tech companies like Anthropic, Google, and Rabbit have all adopted AI agents for their respective platforms, OpenAI only just announced yesterday that it'd be releasing Operator, an agent that can go to the web to perform tasks independently for you.
The Operator is powered by a new model Computer-Using Agent (CUA) which is trained to interact with graphical user interfaces (GUIs), combining GPT-4o's vision capabilities with advanced reasoning through reinforcement learning.
According to OpenAI, Operator automates tasks like flight books, making hotel reservations, shopping online, and so much more. The Operator interface has several task categories where users can choose different automation to perform. You can also save prompts for quick access on the homepage when dealing with repetitive tasks like getting supplies.
Operator will not work like third-party applications needing authorization and API integrations for security reasons, it can simply see and interact with your browser directly and perform tasks or actions easily. According to OpenAI, it'll be fast and compatible with all browsers.
To get started with Operator, simply describe the task you’d like to perform and the Operator will handle it. While this AI agent is meant to automate tasks, its performance is limited when dealing with tasks that require log in, payment details, or deleting certain permanent records. Here, you'll need to manually input those details for security or manually perform the deletion.
Currently, the model has rate limits. OpenAI says that Operator can perform multiple tasks at once, but that it has dynamic limits. It also has a general usage limit which resets daily.
We're certain fans are excited about trying this out, however, OpenAI is starting a small rollout of the Operator to ChatGPT Pro users in the U.S., before expanding to a larger audience.
As OpenAI monitors the release in the U.S., we expect a full rollout soon with changes to those limitations.