OpenAI has launched GPT-6 Cyber, an AI model built to identify security vulnerabilities and help researchers fix them before attackers can exploit them.
OpenAI · Product launch
GPT-6 Cyber: an AI model built to find flaws before attackers do
OpenAI positions GPT-6 Cyber as a defensive tool for security researchers, aimed at spotting software and system weaknesses early enough to be patched.
How the defensive loop is meant to work
The pitch rests on speed: flaws flagged before malicious actors find them.
STEP 1
Model flags flaws
Audits code and infrastructure for vulnerabilities.
→
STEP 2
Researchers review
Defenders confirm and prioritize each finding.
→
STEP 3
Patch before exploit
Fixes ship before attackers can use the flaw.
The defensive case
Shifts the balance of cyber defense toward protectors.
Targets weaknesses early, while they can still be patched.
Part of OpenAI's wider push into specialized professional fields.
The dual-use concern
The same analysis that finds a flaw could help an attacker exploit it.
Access controls, safeguards and independent testing matter here.
Who gets access, and how widely, remains an open question.
What is not yet verifiable
Claims about accuracy or real-world effectiveness cannot yet be independently checked.
Detailed performance benchmarks
Independent evaluations
Availability terms
Pricing
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...