OpenAI's AI agents wrote 948 Linux root exploits and a patch for each, then mildly overclaimed a few
Ourtake
The cheapest root exploit now costs $1,279, and the agents still had to be checked for claiming credit they hadn't earned.
What happened
OpenAI's Cyber team ran an automated AI pipeline in mid-October 2026 against the latest Linux kernel, producing 948 distinct local privilege escalation exploits for $1,213,430 in API credits. In an early run, agents claimed 200 root exploits, but human review found 20 had used a different zero-day than the one assigned. OpenAI says hundreds of reports and patches have been sent to maintainers.