thehackernews.com
Nov 7, 2026
OpenAI Launches GPT-6 Cyber After AI Helps Find 81 Firefox Flaws, 948 Linux Exploits
AI-summarised brief · reviewed before publication
OpenAI unveiled GPT‑6 Cyber, a security‑focused AI model that surpasses prior versions in vulnerability detection and alignment. The model discovered 948 Linux privilege‑escalation exploits, 81 new high‑severity Firefox flaws, and a critical Brotli vulnerability (CVE‑2026‑111497, CVSS 9.2). GPT‑6 Cyber achieves 99.93% compliance with task boundaries and scores 48.5% on the ExploitGym benchmark, up from 32.1% for GPT‑5.6 Cyber. It is initially released to select Daybreak Red customers via API and Codex, with enhanced guardrails to prevent misuse.
💡 Why It Matters
- · By automating the discovery of critical software flaws, GPT‑6 Cyber shifts the balance toward proactive defense, enabling developers to patch vulnerabilities before attackers exploit them.
- · Its superior alignment reduces the risk of malicious misuse, addressing growing concerns over AI‑driven cyber threats.
