OpenAI GPT-6 Cyber uncovers 81 Firefox vulnerabilities
Original title: OpenAI’s GPT-6 Cyber found 81 Firefox flaws earlier AI testing missed
The model identified 81 high-severity Firefox flaws and developed exploits for 948 Linux kernel issues during internal testing.
OpenAI released GPT-6 Cyber, trained to detect vulnerabilities and develop exploits. It found 81 previously unknown Firefox high-severity issues that Mozilla patched and helped develop working exploits for 948 Linux kernel vulnerabilities. In controlled tests, the model enabled agents to escape virtual environments. OpenAI added safeguards like monitoring and sandboxing. The model is available via API to approved customers.
Unconfirmed: Number of 948 Linux kernel vulnerabilities unconfirmed in external reports
- 81 Firefox high-severity vulnerabilities found and patched
- 948 Linux kernel exploits developed
- Outperformed predecessor on ExploitGym with 48.5% score
- Classified as Critical under Preparedness Framework
Impact: unmodified software vulnerabilities discovered in Firefox and Linux kernel
- 1.upgrade to fixed Firefox versions
- 2.review OpenAI API access logs
- Vendors
- OpenAI
- Products
- GPT-6 CyberFirefox
- Organizations
- Mozilla
From the original article
2 min read at helpnetsecurity.com
OpenAI has introduced GPT-6 Cyber, a version of GPT-6 Astra trained to find software vulnerabilities, test how attackers could exploit them, and help develop fixes. The model is designed for authorized security testing of websites and applications. Vulnerability research and security testing OpenAI used GPT-6 Cyber to help secure its systems, including the infrastructure used to train its AI models.
By Anamarija Pogorelec · helpnetsecurity.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.