OpenAI says autonomous security system found 948 Linux kernel flaws and built working exploits

OpenAI disclosed that its autonomous cybersecurity system scanned about 2.8 million lines of Linux kernel code, identified 948 distinct local privilege-escalation vulnerabilities, and produced 1,169 working exploits plus candidate patches. The October run cost roughly $1.21 million in API credits, about $1,279 per vulnerability, and a newer unreleased cyber model found more than five times as many flaws as an earlier version. Human reviewers and maintainers remain responsible for validating fixes.