OpenAI Says Pipeline Found 948 Linux Root Exploits for $1.2 Million
Summary
OpenAI says its autonomous pipeline, built on an upcoming cyber model, reproduced 1,724 Linux kernel vulnerabilities and produced 948 local privilege escalation exploits for $1,213,430 in API credits. OpenAI says its agents used at least 10 exploitation techniques that were new to outside kernel experts. Maintainers patched 61 of the LPEs during the 10-day scan.
Key Points
- Scout agents narrow the audit to 5,034 targets, about 2.8 million lines or 7% of the kernel, OpenAI says.
- In a July run, OpenAI says GPT-5.6 Sol found 166 unique LPEs for $513,690 in API credits.
- OpenAI reports 21 LPE exploits against a kernelCTF kernel for $32,936 in credits, versus 26 accepted 0-days in 2026.