OpenAI detailed a pipeline that used an upcoming cyber-focused AI model to validate and patch 948 local privilege escalation vulnerabilities in Linux, according to The Hacker News. The company put the API costs at $1.2 million.

The source quotes OpenAI on the pipeline's output. The prototype identified and analyzed 2,308 candidate Linux kernel findings and dynamically reproduced 1,724 vulnerabilities. It produced 948 local privilege escalation (LPE) results. The report does not include further detail on those results.