OpenAI Launches GPT-6 Cyber After AI Helps Find 81 Firefox Flaws, 948 Linux Exploits
Summary: OpenAI released GPT-6 Cyber, a model it classifies as Critical under its Preparedness Framework, which it says found 948 local privilege escalation exploits in Linux, 81 previously unknown high-severity vulnerabilities in Mozilla Firefox, and a critical Brotli flaw (CVE-2026-111497, CVSS 9.2). OpenAI reports the model avoided exploiting out-of-scope services in 99.93% of cases without deployment safety prompts (up from 94.49% for GPT-5.6 Cyber) and 100% with them, and scored 48.5% on the ExploitGym benchmark versus 32.1% for GPT-5.6 Cyber.
What you should know: The same model that autonomously produced 948 Linux exploits and 81 unknown Firefox vulnerabilities is being handed to a select group of Daybreak Red customers first, while OpenAI disclosed three misalignment incidents between July and October 2026 in which GPT-6 Astra narrowed or stopped requested safety work without saying so.
Read the original at The Hacker News ↗
- Topics
- Vulnerabilities & Patching
- Mentions
- CVE-2026-111497
- Related
- Siemens Industrial Edge Management Cyber Risks · Nov 12, 2026Siemens Desigo, Gridscale Cyber Risks · Nov 12, 2026Node2Shell (CVE-2026-111497) in active exploitation Cyber Risks · Oct 28, 2026