← Back

Wire · opportunities

OpenAI launches GPT-6 Cyber, which found 81 high-severity Firefox flaws

Published

5 November 2026

Topic

opportunities

◆ Sectors

Cybersecurity

◆ Source

Read at marketscale.com →

◆ Verified

Fusion42 · 6 November 2026 · Fusion42 review

OpenAI launched GPT-6 Cyber, a specialised GPT-6 Astra model trained for authorised vulnerability research, which identified 81 new high-severity Firefox vulnerabilities and numerous Linux kernel exploits, with many fixes already deployed. Access to the model is limited to approved Daybreak Red customers via OpenAI's API and Codex.

This Wire brief sits within Fusion42's coverage of Cybersecurity, and 3 sources have reported it.

◆ ◆ The Wire takeaway

You now face a new competitor that finds vulnerabilities faster and more accurately than before. If your cybersecurity tool relies on manual testing, this is the client threat and integration opportunity you must deal with immediately.

◆ Coverage

3 sources · 5 Nov 2026

◆ Related on Wire

◆ Topics

Cybersecurityopenaigpt-6-cybervulnerability-researchfirefox-flawslinux-kernelapi-access