On Friday, OpenAI commenced the deployment of its AI agent, dubbed Operator, in multiple countries. Initially made available in preview to ChatGPT Pro users in the United States since its introduction in January, the Operator is now being utilized in a broader range of major regions. However, European nations will be excluded from this rollout for the time being.
OpenAI’s Operator AI Agent Officially Launches
The announcement was shared via a post on X (formerly Twitter) from OpenAI’s official account, which revealed that Operator is now accessible to ChatGPT Pro users in countries including Australia, Brazil, Canada, India, Japan, Singapore, South Korea, and the UK, alongside existing users in the US.
Nevertheless, OpenAI indicated that users in Europe will need to wait longer for access to the AI agent. The post noted that the team is still in the process of enabling the Operator for users in the EU, Switzerland, Norway, Liechtenstein, and Iceland.
When Operator was first unveiled, OpenAI detailed that it operates using a specialized AI model known as the Computer-Using Agent (CUA). This model combines computer vision capabilities derived from GPT-4o with advanced reasoning features potentially sourced from an undisclosed OpenAI model, which might be the o3 model. The AI agent has undergone post-training through reinforcement learning techniques.
The CUA is designed with agentic abilities, allowing it to interact with graphical user interfaces (GUIs) that include buttons, menus, and text fields on the screen. It features a dedicated browser, enabling it to carry out actions in a separate digital environment and providing users with the flexibility to focus on other tasks.
Operator is capable of processing both text and image inputs. While executing tasks, it analyzes the raw pixel data displayed on the screen and employs a virtual keyboard and mouse within its sandboxed environment. The AI agent is structured to handle multi-step tasks, navigate errors, and adapt to unforeseen circumstances effectively.
