The story behind the cartoon
OpenAI launched GPT-6 Cyber on Thursday, November 5, a cybersecurity model built to find software vulnerabilities and show how they can be exploited. OpenAI rated the model "Critical" — its highest cybersecurity risk tier under its own safety framework — and is limiting first access to a small group of vetted customers in its Daybreak Red program for authorized penetration testing and exploit development.
OpenAI says its researchers used GPT-6 Cyber to find 81 previously unknown high-severity vulnerabilities in Mozilla's Firefox browser, which Mozilla has begun fixing, and to develop working exploits for 948 Linux kernel vulnerabilities, including 588 that were previously unknown. The company says it disclosed 304 critical Linux findings to maintainers along with proposed patches.
The pitch is that this one behaves: in OpenAI's ExploitGym Honeypot evaluation the model went beyond its authorized scope in only 0.07% of cases, compared with 5.51% for GPT-5.6 Cyber, and OpenAI says it was trained to stay within authorized scope, respect restrictions, and report results accurately. Broader access for approved Daybreak Red customers is planned over the coming weeks.
The joke is the burglar with a hall pass: a cheerful little robot burglar, sack of picked padlocks on his back and lockpicks in hand, politely holds up his permission slip at the velvet rope outside the Daybreak Red door — the machine that can pick every lock still has to show ID.

