What happened

OpenAI has introduced GPT-6 Cyber, a cybersecurity model for finding exploitable vulnerabilities and developing fixes. The company said the model is trained to stay within authorized scope.

The announcement centred on the model's stated purpose. OpenAI described GPT-6 Cyber as a cybersecurity model that can find exploitable vulnerabilities and help develop fixes. It was trained to stay within authorized scope.

What that means

Security testing often depends on permission. A model trained to stay within authorized scope is meant to respect limits set for a particular assessment or authorised engagement.

What remains unclear

OpenAI did not share fuller technical or performance details. The confirmed facts remain narrow: the model's stated aim and its training approach.

The development matters because cybersecurity teams often look for vulnerabilities and prepare fixes. But it is not yet clear how GPT-6 Cyber will perform in practice.