OpenAI's GPT-6 Cyber Finds 948 Linux Privilege Escalation Exploits
OpenAI ran GPT-6 Cyber, its cybersecurity-focused model, across the Linux kernel: 2,308 vulnerability candidates, 1,724 reproduced, and 948 working local privilege escalation exploits. The full scan cost $1,213,430 in API fees, or $1,279 per exploit.
1 source 6m ago
Reporting from 1 source: GIGAZINE.
Why it matters
The $1,279-per-exploit figure puts a hard price on automated bug hunting and gives maintainers a patching backlog measured in hundreds of fixes, which is why OpenAI argues automatic patch generation is now the bottleneck.
Key Facts
- OpenAI used its GPT-6 Cyber model to scan the Linux kernel for security vulnerabilities.
- The scan surfaced 2,308 candidates, reproduced 1,724, and produced 948 working local privilege escalation exploits.
- The full Linux verification cost $1,213,430 in API fees, or $1,279 per exploit.
- OpenAI reported 304 critical vulnerabilities to Linux maintainers with candidate patches, per community guidance.
- GPT-6 Cyber also produced reproducible tests for 81 new high-severity Firefox vulnerabilities disclosed to Mozilla.
Following Linux community guidance, OpenAI reported 304 critical vulnerabilities to maintainers with candidate patches and is working on the rest.
OpenAI has attached a price to AI-assisted kernel auditing: $1,213,430 in API usage for the Linux run, against exploits it values at $1,279 each.
The same model was also turned on Firefox. It produced reproducible tests for 81 new vulnerabilities Mozilla rates high severity and found 30 known issues to be more severe than previously reported, and Mozilla began applying fixes shortly after the report. For Linux, patches for more than 100 exploits were applied or in preparation during the 10-day verification, with about 300 more under public discussion.
Written with AI by Yomimono from the 1 cited source below, including Japanese-language reporting where cited. See our editorial policy.
Add Yomimono as a preferred source on Google