OpenAI's new model automatically uncovers 948 Linux kernel privilege escalation vulnerabilities, with each exploit costing only $1,279

On November 4 local time, OpenAI released a security research report, stating that with the help of its upcoming brand-new cybersecurity model, it has built a complete closed loop covering automated vulnerability discovery, validation, exploitation, and remediation, and delivered impressive results in real-world tests. A real-world test in October 2026 showed that the system could dynamically reproduce and exploit vulnerabilities from among thousands of candidates, demonstrating formidable automated security attack and defense capabilities; the related tests also involved massive samples.