OpenAI is upgrading its AI agent, Operator, to utilize its advanced o3 reasoning model. Previously powered by a custom GPT-4o model, the new o3 Operator boasts improved math and reasoning capabilities. OpenAI emphasizes that o3 Operator, while enhanced, maintains a multi-layered safety approach and lacks direct access to coding environments. This upgrade positions Operator among other sophisticated AI agents from companies like Google and Anthropic, all vying to create reliable, unsupervised task-completion tools.

Prepared by Jonathan Pierce and reviewed by editorial team.