OpenAI GPT-6 Cyber Rated Critical: Firefox and Linux Impact Explained

[jiaolixan]

OpenAI has introduced GPT-6 Cyber, a security-focused model designed to identify exploitable software vulnerabilities and build working exploit code. Under OpenAI’s Preparedness Framework, the model’s cybersecurity risk is rated “Critical”—the highest tier in the framework. Its predecessor, GPT-5.6 Cyber, was rated “High.”

The model is not part of ChatGPT. GPT-6 Cyber is available through the API and Codex to a select group of approved clients in OpenAI’s “Daybreak Red” program. That program is intended for authorized offensive security work, such as penetration testing. OpenAI says access will expand to more Daybreak Red clients over time.

Why OpenAI Rated GPT-6 Cyber as Critical

OpenAI’s benchmark results help explain the tight controls. In the ExploitGym test, which asks a model to exploit specific vulnerabilities on a remote system, GPT-6 Cyber achieved a 48.5% success rate, up from 32.1% for its predecessor.

OpenAI also says the model “avoided exploiting out-of-scope services in 99.93% of tests” when no specific deployment prompts were provided. The company declined to share full details of the remaining testing with The Register.

Access controls include IP address whitelisting by default, monitoring for compromised credentials, and blocks on attacks targeting critical infrastructure such as power plants, water treatment facilities, and industrial control systems.

Firefox and Linux Users Are Already Seeing the Impact

OpenAI says its researchers used GPT-6 Cyber to uncover 81 previously unknown, high-severity vulnerabilities in Firefox. Some have already been patched. In a security advisory for Firefox 159 released on October 27, Mozilla noted that OpenAI had addressed 41 of 167 identified vulnerabilities, all rated “high” severity.
The current Firefox version is 159.0.1, with Extended Support Releases at 153.6 and 115.44. Users who disabled automatic updates should check “Help” > “About Firefox.”

The model also reproduced 1,724 vulnerabilities in the Linux kernel and generated functional exploit code for 948 of them—code that could allow an attacker to gain root access. OpenAI reported 304 critical vulnerabilities to Linux kernel maintainers and submitted draft patches. Because distributions such as Ubuntu and Fedora need time to roll out fixes, users should install kernel updates promptly in the coming weeks.

Access Is Limited to Vetted Security Teams

GPT-6 Cyber is not a general-purpose chatbot upgrade. It is a controlled security tool for approved teams, and OpenAI’s safeguards are designed to keep it away from critical infrastructure targets. The debate now is whether limiting such models to vetted internal teams is enough to protect digital infrastructure.

What This Means for OpenAI Account Shoppers

The model cannot be reached through ordinary ChatGPT Pro Accounts. OpenAI says GPT-6 Cyber is separate from ChatGPT and is available only through approved API and Codex access.
Some users searching for a Cheap OpenAI Account may see GPT-6 Cyber headlines and assume a consumer subscription unlocks the model. It does not. Daybreak Red approval, IP whitelisting, and OpenAI’s security controls are separate from consumer account plans.

Searches for “Buy Cheap OpenAI Account” may lead to third-party listings, but those listings do not provide access to GPT-6 Cyber or the Daybreak Red program.
Marketplaces such as z2u.com may list account options for general OpenAI tools, but they are not a route to GPT-6 Cyber. Access to the security model remains restricted to vetted teams approved by OpenAI.

FAQ

Is GPT-6 Cyber available in ChatGPT?
No. OpenAI says GPT-6 Cyber is available via API and Codex to approved Daybreak Red clients.

Why is GPT-6 Cyber rated Critical?
OpenAI’s Preparedness Framework places its cybersecurity risk at the highest tier because it can identify vulnerabilities and construct working exploit code.

What should Firefox and Linux users do?
Keep browsers and kernels updated. Firefox users should check their version, and Linux users should install kernel updates as distributions release them.

openai3.jpg