threatdeck
Live 14:45 UTC ·

Model Watch Medium For CISO, AppSec, AI Eng, Red Team AI-assisted

OpenAI raises GPT-6 Cyber risk rating to Critical after Linux, Firefox findings

OpenAI has introduced GPT-6 Cyber, a model for security research, and has raised its cybersecurity risk rating from High to Critical. OpenAI reports that the model reproduced 1,724 Linux kernel vulnerabilities, developed working exploits for 948, and produced reproducible tests for 81 high-severity Firefox flaws. The source states these are OpenAI's own results and should be treated as claims to check.

Why it matters

Exploit development for kernel and browser flaws is becoming faster and cheaper, so organizations running Linux, Firefox, PHP, or Node.js may face more disclosed flaws to patch. Dragos says the skill floor for reconnaissance and vulnerability research is dropping.

Do this

  1. Confirm Firefox and Linux kernel builds in your estate include the upstream fixes for the disclosed issues, and track maintainer advisories for the 304 reported critical findings.
  2. Prioritize patching for PHP, Node.js and Brotli dependencies, since the source describes critical flaws in each.
  3. If your teams use AI models for vulnerability research, require sandboxing, action review, and human oversight for high-risk work, as OpenAI recommends.
  4. Treat vendor benchmark and alignment figures as unverified until independent evaluations are published; do not change risk decisions on these numbers alone.

Signals

  • NoCISA KEV
  • 22/100Priority
  • 60%Confidence

Evidence

“We’re introducing GPT‑6 Cyber, a major leap in security research: more capable at finding and fixing vulnerabilities, and far more reliable staying within the boundaries researchers set.”

Security Affairs

“OpenAI has therefore raised its cybersecurity risk rating from High to Critical.”

Security Affairs

“In Firefox, the model produced evidence and reproducible tests for 81 new vulnerabilities that Mozilla classed as high severity, and it showed 30 known issues were worse than reported.”

Security Affairs

“The model avoided unauthorized actions in 99.93% of cases, compared with 94.49% for GPT-5.6 Cyber.”

Security Affairs

Sources

Affected / related

GPT-6 Cyber, GPT-6 Astra, GPT-5.6 Cyber, OpenAI, Firefox, Linux kernel, Dragos

Published 2026-11-06 12:26 UTC · updated 2026-11-08 04:28 UTC · AI-assisted summary, checked against quoted sources · How we score

Following

Follow the vendors, products, AI tools, techniques and actors you care about. Matching stories, CVEs and exploited bugs are marked, a Following tab appears in the feed, and returning visits show what changed. Saved only in this browser.

Presets

Get updates

A personal RSS feed of new stories on everything you follow. Paste it into any feed reader, Slack or Teams RSS app.