GPT-6 Cyber Finds 81 Firefox Flaws, 948 Linux Exploits — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI unveiled GPT-6 Cyber on Thursday and is shipping it to a limited group of Daybreak Red users.
- GPT-6 Cyber combines stronger cybersecurity capabilities with what OpenAI described as a “step-change improvement” in model alignment.
- OpenAI said GPT-6 Cyber is more capable at finding and fixing vulnerabilities.
- OpenAI reported that GPT-6 Cyber found 81 previously unknown Firefox flaws and 948 Linux root-escalation exploits in 30 hours.
- GPT-6 Cyber recorded a 48.5% score on ExploitGym.
- OpenAI pledged $1 million to support security research.
Why it matters: The limited Daybreak Red cohort gains access to a model that reported 81 previously unknown Firefox flaws and 948 Linux root-escalation exploits in 30 hours. OpenAI’s $1 million research pledge puts money behind security work, while the alignment claim makes model behavior—not only vulnerability discovery—a stated measure of the release.
Ask SkimNews
