OpenAI has commenced the rollout of ChatGPT Agent, a powerful new tool designed to transform its chatbot into an autonomous digital assistant. Originally announced on July 18, the feature is now being made available to all ChatGPT Plus, Pro, and Team subscribers. This development marks a significant shift from a conversational model to an "agentic" system capable of executing complex, multi-step tasks on a user's behalf.
The ChatGPT Agent carries out tasks using its own virtual computer, fluidly shifting between reasoning and action. It combines the abilities of OpenAI's older tools, "Operator" and "Deep Research," to interact with websites, synthesize information, and leverage ChatGPT's intelligence. Its capabilities include the ability to analyze calendars, plan and buy ingredients for recipes, or create slide decks after analyzing competitors. To achieve this, the Agent is equipped with a suite of tools, including a visual browser, a text-based browser, a terminal for code execution, and direct API access.
Users can activate Agent mode from the tools menu or by typing /agent. Once enabled, the user describes the task, and the Agent begins executing it, pausing to request clarification or confirmation when needed. A key feature is that the user is always in control, with the ability to interrupt, take over the browser, or stop tasks at any point. Strong safety features include user permission prompts, real-time monitoring, and refusal of risky actions to protect privacy.
Access to this new capability differs by subscription plan. ChatGPT Plus users receive 40 Agent messages per month, while Pro users have a higher limit of 400 messages. This difference has led some Plus users to feel constrained, although each message can initiate a session lasting 5-30 minutes to complete a complex task. OpenAI plans to expand access to Enterprise and Edu users in the coming weeks.
The introduction of ChatGPT Agent represents a major step toward more advanced and capable AI assistants. By combining research, web interaction, and task execution, the Agent can automate workflows, create spreadsheets and presentations, and assist with competitor analysis. While the rollout experienced some initial delays, OpenAI emphasizes that this launch is just the beginning, with regular improvements expected to make the tool even more capable over time.