OpenAI Announces Enhanced Accuracy of Operator Agent with O3 Update

OpenAI has announced that the Operator Agent has been upgraded to utilize the o3 model, resulting in enhanced reasoning capabilities. Operator, which is currently in a research preview phase, empowers users to delegate various web tasks to an AI agent, thereby streamlining routine activities such as form completion and product ordering.

Since its introduction in January 2025, Operator has only been accessible to subscribers of the $200 Pro and Enterprise tiers. This agent has now undergone significant improvements, aligning with OpenAI’s objective to evolve Operator into a highly efficient tool.

The previous version of the Agent operated on the 4o model, which lacked advanced reasoning capabilities. The incorporation of the o3 model marks a transition to a more robust and accurate interaction with web environments. OpenAI reports that, with this upgrade, Operator exhibits increased persistence and precision, resulting in an improved overall task success rate. The responses generated are now clearer, more comprehensive, and better organized.

Despite these advancements, access to Operator remains limited to ChatGPT Pro users on a global scale. Currently, this upgraded model is exclusively available within the Operator feature of ChatGPT. As a point of consideration, some potential users may question the value of investing $200 for automated tasks such as grocery shopping. Consequently, OpenAI is internally exploring methods to enhance the utility of this tool.

Looking forward, OpenAI intends to make Operator available to $20 Plus subscribers in the near future, potentially broadening the tool’s availability and usefulness.