What Changed
OpenAI has released GPT-5.6-Cyber, a version of GPT-5.6 Sol trained for vulnerability research, penetration testing and incident response, with fewer refusals on higher-risk dual-use tasks such as finding zero-days and building exploit chains. It is available through Daybreak Red, a new access tier for vetted organizations doing authorized security testing. It succeeds GPT-5.5-Cyber, released in June.
On OpenAI’s internal Advanced Cybersecurity Completion Rate evaluation, GPT-5.6-Cyber completed 95.0% of requests, compared with 1.5% for GPT-5.6 Sol, 2.0% under Daybreak Blue access and 57.3% for GPT-5.5-Cyber.
What The Model Has Found
OpenAI credits the model with CVE-2026-15903 (CVSS 8.8), an out-of-bounds read and write in Chrome’s V8 engine that Google patched in mid-July, which could be chained with a second flaw the model found to escape the V8 heap sandbox. OpenAI also cites at least five flaws in a popular mobile operating system, three critical bugs in a popular database and more than 400 privilege-escalation issues in a popular kernel.
Know what matters before your first meeting.
Weekday mornings. Five minutes. What changed in security, why it matters.
By subscribing you agree to our Privacy Policy.
Free. Weekday mornings. 5 minutes or less.
OpenAI notes the model does worse than GPT-5.6 Sol on open-ended repository bug hunts because its vulnerability reports tend to be shorter and less detailed.
Why This Is An Operations Problem
Early partners include Accenture, Akamai, Cisco, Cloudflare, CrowdStrike, Fortinet, IBM, Palo Alto Networks, PwC and Sophos. More vendors with these tools means more vulnerability reports, and more patches for customers to absorb.
Fixing is still harder than finding. 1Password research cited in the report found LLM-generated patches fully fixed the bug without changing behavior only 26.0% of the time.
What To Verify First
- Make sure patch processes can handle higher volumes of vendor advisories.
- If you trial AI-assisted testing, define scope, approvals and logging before granting access.
- Have humans review AI-generated fixes before merging.
What Teams May Be Underestimating
OpenAI itself says that models with reduced safeguards carry risks “whether from misuse or misalignment.” Access controls on these tools deserve the same scrutiny as any other privileged security capability.
Know what matters before your first meeting.
Weekday mornings. Five minutes. What changed in security, why it matters.
By subscribing you agree to our Privacy Policy.
Free. Weekday mornings. 5 minutes or less.
Source Context
This item was rebuilt from the following primary sources and original reporting:
Related In The Daily Brief
See this item in The 5-Minute Cyber Brief
Newer CyberExperts coverage on this topic
This article still works as background. If you want the current picture, start with the freshest related coverage below and today's brief.
Coder’s registry infrastructure compromised to push malicious modules
The Coder incident is a supply-chain lesson in miniature: once attackers can tamper with trusted module distribution, defenders are no longer verifying...
Attackers Turn Trusted Node.js Runtime Into Malware Delivery Tool in Targeted Attacks
The useful lesson in this campaign is not that Node.js is bad. It is that adversaries keep choosing legitimate runtimes defenders already...
HPE patches critical ArubaOS-CX remote code execution flaw
ArubaOS-CX deserves attention because switching software rarely gets treated with the same urgency as identity or edge security until exploitation arrives. A...
Monday’s brief: GlobalProtect logins without passwords, then SonicWall, AhsayCBS and NetScaler
The fastest way to catch up on what changed after this article was published.
Start your morning with the signal that matters.
Get the biggest cybersecurity developments, why they matter, and where to go deeper on CyberExperts.
By subscribing you agree to our Privacy Policy.
Free. Weekdays. Built for operators.
