Technology & Science
OpenAI Rolls Out ‘GPT-6 Cyber’ to Vetted Daybreak Red Users with “Critical” Risk Rating
On 5 Nov 2026 OpenAI began limited deployment of GPT-6 Cyber—its first AI model rated “Critical” for cybersecurity—granting select Daybreak Red customers API access after internal tests showed a five-fold drop in unintended hacking and major gains in exploit discovery.
Focusing Facts
- Internal red-team trials found GPT-6 Cyber uncovered 81 new high-severity Firefox bugs and built working exploits for 948 Linux-kernel flaws (588 previously unknown).
- In the ExploitGym Honeypot test the model strayed out-of-scope <1 % of runs, down from ~5 % with GPT-5.6 Cyber.
- OpenAI disclosed 304 critical Linux findings with proposed patches and estimates testing the model cost about US$1.2 million in API credits.
See how 3 sources reported this story.
- ✓ Full multi-perspective analysis on every story
- ✓ Primary source links for every claim
- ✓ Daily email briefing — no algorithm
Perspectives in this article
- General tech news outlets
- Cybersecurity specialist media
- Business and financial press