OpenAI releases GPT-6 Cyber model for vulnerability hunting
Original title: OpenAIs GPT-6 Cyber can both break into systems and propose how to fix them
GPT-6 Cyber finds vulnerabilities in Firefox and Linux kernel with improved safety; initial access limited to select customers.
OpenAI introduced GPT-6 Cyber, an AI model trained to find security vulnerabilities, build exploits and suggest patches. The model achieved 99.93 percent safety compliance without instructions and scored 48.5 percent on ExploitGym. It reported 81 new severe vulnerabilities in Firefox and reproduced 1,724 in the Linux kernel. Access is initially limited to a select group of customers.
Unconfirmed: No independent testing of the model has been published so far.
- GPT-6 Cyber found 81 severe vulnerabilities in Firefox
- Reproduced 1,724 vulnerabilities in Linux kernel with 948 working exploits
- Campaign cost $1,213,430 or $1,279 per exploit
- Initial access limited to select Daybreak Red customers
- Vendors
- OpenAI
- Products
- GPT-5.6 Cyber
- Organizations
- MozillaLinux maintainers
- AI models
- GPT-6 CyberGPT-6 Astra
From the original article
4 min read at the-decoder.com
With GPT-6 Cyber, OpenAI is introducing an AI model built to find security vulnerabilities in software, develop exploits, and suggest patches. According to OpenAI, it has already found 81 previously unknown vulnerabilities in Firefox and developed 948 working exploits in the Linux kernel, significantly more than its predecessor.
By Matthias Bastian · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.