OpenAI has introduced GPT-6 Cyber, a version of GPT-6 Astra trained to find software vulnerabilities, test how attackers could exploit them, and help develop fixes. According to the report, the model is designed for authorized security testing of websites and applications.
The report says OpenAI used GPT-6 Cyber to help secure its own systems, including the infrastructure used to train its AI models. It also says that in controlled security tests, agents using GPT-6 Cyber and other OpenAI models combined vulnerabilities to, but the article text is cut off at that point.
The headline states that GPT-6 Cyber found 81 Firefox flaws that earlier AI testing missed. The supplied text does not include details on those flaws, the testing process, or the comparison.