San Francisco, 6 November 2026 – OpenAI released GPT-6 Cyber to a limited group of vetted users on 5 November, giving authorised security teams access to a model built to find and validate software vulnerabilities. The company classified it as Critical for cybersecurity under its own Preparedness Framework. It said the model helped identify 81 previously unknown flaws that Mozilla classified as high severity in Firefox.
For Asian companies that build or depend on widely used software, faster vulnerability discovery could shorten the interval between finding a weakness and fixing it. But the announcement also illustrates why access is restricted. The same ability to establish that a flaw can be exploited makes a model valuable to defenders and sensitive to misuse. Neither the rating nor the reported results should be mistaken for an independent assurance of safety.
Unlock the Full Article
This article is exclusive to The Ledger Asia Subsribers / PAID members.
Already have an account? Log in here

