OpenAI is reportedly preparing to release a new AI tool called Operator, which can autonomously handle tasks like writing code and booking travel. According to software engineer Tibor Blaho, who claims to have uncovered evidence of the tool, it will be available on macOS.
Blaho discovered hidden options in OpenAI’s ChatGPT client that allow users to define shortcuts for toggling and quitting Operator. Additionally, the company’s website contains references to Operator, including comparison tables with other AI systems.
While the performance of Operator is not yet publicly disclosed, Blaho shared leaked benchmarks suggesting it may not be 100% reliable. The tool scores well on some tasks but struggles with others, such as creating a Bitcoin wallet or signing up for a cloud provider.
OpenAI’s imminent release of Operator comes as rival companies like Anthropic and Google are also entering the AI agent space. The market for AI agents is expected to reach $47.1 billion by 2030, according to analytics firm Markets and Markets.
The development of Operators has raised concerns about their safety, with some experts warning that the technology could become increasingly speculative and risky. OpenAI co-founder Wojciech Zaremba recently criticized Anthropic for releasing an agent without adequate safety measures, highlighting the need for careful consideration in the release of AI tools like Operator.
Source: https://techcrunch.com/2025/01/20/openais-agent-tool-may-be-nearing-release