1 source · latest 28m

OpenAI launches restricted GPT-6 Cyber model for finding and testing software flaws

AI SECURITYFirefoxredteam/daily · 1 source

Severity: Medium
Sources: 1

What happened

OpenAI reports GPT-6 Cyber found 81 previously unknown high-severity Firefox vulnerabilities, with Mozilla now releasing fixes, and produced working exploits for 948 Linux kernel vulnerabilities (588 unknown), with candidate patches going to maintainers. In a sample of 166 exploits, 10 techniques were described as novel and potentially enabling local root. Out-of-scope actions were 0.07% in the ExploitGym Honeypot evaluation versus 5.51% for GPT-5.6 Cyber. No CVE IDs are given; defenders should apply Mozilla fixes as released and track the OpenAI Patch the Planet dashboard, which reportedly shows under 25% remediation.

For security analysts · Written by AI from 1 source · Wrong?

What to do

Update Firefox to Mozilla's fixed releases as they become available.
Apply Linux kernel patches once maintainers have reviewed and released them.
Prioritise remediation of known vulnerabilities, as the OpenAI Patch the Planet dashboard reportedly shows under 25% remediation.

Written by AI from the sources, checked against them

Read the full story at the source