OpenAI's New GPT-6 Cyber AI Finds 81 Firefox Flaws, 948 Linux Bugs
The model also generated proof-of-concept exploits and candidate patches, and OpenAI said the run cost about $1.21 million in API credits.
- On Thursday, OpenAI released GPT-6 Cyber, its latest cybersecurity-tuned model, which the company says combines "stronger cybersecurity capabilities with better adherence to researchers' rules of engagement."
- OpenAI designed the model to "stay within authorized scope, respect restrictions and report results accurately," with testing showing it avoided exploiting out-of-scope services over 99 percent of the time during ExploitGym evaluations.
- Researchers used the model to identify 81 previously unknown high-severity vulnerabilities in Firefox and developed exploits for 948 Linux kernel vulnerabilities, including 588 previously unknown bugs.
- Initially available through the API and Codex for select Daybreak Red members, OpenAI plans to expand access to all Daybreak Red customers in the coming weeks.
- Classified as "Critical" under OpenAI's Preparedness Framework, the model includes new checks to block "high-risk activity," such as attacks on critical infrastructure like power plants and water treatment facilities.
142 Articles
142 Articles
Open A.I. launched the GPS-6 Cyber model to enhance capabilities to detect security gaps and test their potential to be exploited, and to help researchers address weaknesses before being employed in cyberattacks.
OpenAI has introduced a new artificial intelligence model, GPT-6 Cyber, which is designed specifically for cybersecurity and vulnerability detection in computer systems. The new model has demonstrated advanced capabilities by identifying nearly 1,000 security vulnerabilities in Linux, as well as dozens of previously unknown problems in the Firefox browser. According to OpenAI, GPT-6 Cyber has been specially trained to detect security vulnerabili…
OpenAI's new GPT-6 Cyber AI finds 81 Firefox flaws, 948 Linux bugs
OpenAI has revealed GPT-6 Cyber, a new artificial intelligence model that has discovered 81 previously unknown high-severity Firefox vulnerabilities and developed working exploits for 948 distinct Linux kernel flaws. [...]
GPT-6 Cyber is finding security flaws before hackers can exploit them
From Firefox to the Linux kernel, OpenAI's new cybersecurity model is uncovering serious flaws that earlier AI systems missed and helping developers fix them before attackers can exploit them.
Coverage Details
Bias Distribution
- 48% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
