OpenAI’s AI agent ‘Operator’ likely to be released in January

OpenAI is reportedly working on an artificial intelligence (AI) agent named “Operator” that can take over a user's PC and perform tasks autonomously.

The existence of OpenAI’s AI-driven agent was first reported by Bloomberg, and now a reliable AI industry leaker and software engineer named Tibor Blaho claims to have uncovered evidence of the tool's development.

OpenAI's Operator to release in January

According to Blaho, OpenAI's ChatGPT client for macOS has gained hidden options to define shortcuts for "Toggle Operator" and "Force Quit Operator."

Although these are not yet publicly visible, OpenAI's website contains references to Operator.

The AI agent tool is capable of handling complex tasks such as writing code and booking travel. However, leaked benchmarks suggest that the Operator might not be 100% reliable, depending on the task.

On OSWorld, a benchmarking site, Operator scored 38.1%, ahead of Anthropic’s computer-controlling model but fell short of 72.4% human score.

While OpenAI has not officially confirmed the Operator’s release date, the company is targeting January as the release month, according to The Information.

If the claim has anything to go by, this would mark OpenAI’s significant entry into the AI agent space to compete with rivals such as Anthropic, Google, and others.

Notably, the market for AI agents is expected to grow significantly, with analytics firm Markets and Markets predicting that it could be worth $47.1 billion by 2030.