OpenAI AI Pipeline Finds 948 Linux Root Exploits for $1.2 Million

In a landmark—and sobering—demonstration of AI-powered cybersecurity research, OpenAI has revealed that an autonomous pipeline built on an unreleased cyber model discovered and exploited 948 unique local privilege-escalation (LPE) vulnerabilities in the Linux kernel during a single 30-hour run in mid-October 2026.

The pipeline analyzed the Linux kernel’s more than 40 million lines of code, narrowing its audit to roughly 2.8 million lines across 5,034 targets. It then:

• Triaged 2,308 candidate findings
• Dynamically reproduced 1,724 vulnerabilities
• Produced 948 working root exploits, each paired with a candidate patch

The run cost $1,213,430 in API usage, or roughly $1,279 per unique flaw.

For perspective, hacking competitions such as Pwn2Own and TyphoonPWN have historically paid between $20,000 and $70,000 for a Linux kernel privilege-escalation exploit. OpenAI estimates that the 948 exploits could be worth between $18.9 million and $66.3 million on the open bounty market.

OpenAI says its goal is defensive. Through its Patch the Planet initiative, the company has begun sending vulnerability reports and patches to the Linux community, starting with the most critical flaws. The model behind the pipeline has not yet been publicly released.

Source: Help Net Security: OpenAI’s upcoming AI model built 948 working Linux root exploits in a single run