// AI news · 2026-11-06 · The Next Web
OpenAI Releases GPT-6 Cyber, Finds 81 Firefox Flaws
OpenAI released GPT-6 Cyber, a security model that outperforms earlier versions in flaw detection and boundary adherence.
Published 2026-11-06 01:24 (Beijing time)Source The Next WebAuto-summarised, fact-checked against source
01Summary
It achieved 99.93% compliance with restricted service access in tests, surpassing GPT-5.6 Cyber's 94.49%. The model identified 81 high-severity Firefox flaws and reproduced 1,724 Linux kernel vulnerabilities, building exploits for 948. OpenAI disclosed 304 critical Linux issues with proposed patches.
Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.
02Key points
- GPT-6 Cyber achieved 99.93% compliance in restricted access tests
- Found 81 high-severity Firefox flaws and 1,724 Linux kernel vulnerabilities
- Disclosed 304 critical Linux issues with patch suggestions
- Completed 95% of remote attack tasks on new benchmark
03Source
OpenAI’s GPT-6 Cyber found 81 new Firefox flaws and is rated ‘Critical’
The Next Web · published 2026-11-06 01:24 Beijing time
05Related news
- Same source · Cat Wu Leaves Anthropic to Join OpenAI2026-11-05 · The Next Web
- Same source · DeepSeek Open-Sources Super-Agent Coding Agent2026-11-01 · The Next Web
- Same topic · OpenAI Launches GPT-6 Cyber for Cybersecurity2026-11-06 · The Verge
- Same topic · Hugging Face CEO says P(doom) is near zero2026-11-05 · WIRED
- Same day · Nasdaq on track for worst day in 18 months2026-11-06 · Guardian Tech
From AIMap AI news: collected from 1,100+ official and media sources. Summaries only restate the original article; follow the link for the full story. All stories of 2026-11-06 · Archive by date · All AI news · Today