GPT-6 Cyber reportedly rated Critical, surfaces ~1,000 Linux kernel vulnerabilities
sven_ai · x · 2026-11-06
Unverified reports say OpenAI's newly released GPT-6 Cyber earned the top 'Critical' rating for cybersecurity capability.
- Researchers allegedly built an automated system with it that found nearly 1,000 security issues in the Linux kernel, with attempted validation and fixes
- A shared chart shows an Advanced Cybersecurity Completion Rate near 95%, far above GPT-5.5-Cyber's 57%
The takeaway: AI-driven vulnerability discovery is accelerating fast, and the bottleneck may soon shift to whether security teams can patch as fast as AI finds bugs.
Related event: Leaked GPT-6 Cyber Security Model Scores 48.5% on ExploitGym(3 posts)→
More from Models
- Pro user says ChatGPT $200 plan weekly quota drained in 6 hours, suspecting another silent cut — iamrobotbear · 2026-11-06
- Dev hopes for a 397B-A17B model but doubts it will arrive anytime soon — gnukeith · 2026-11-06
- Qwen 4 may launch today and be a 'significant leap forward', claims investor — bindureddy · 2026-11-06
- Reverse-engineering how Google builds Gemini grounding snippets: extractive, lead-biased, noisy — dejanseo · 2026-11-06
- Grok's speech-to-text nails transcription even with blaring music, user reports — XFreeze · 2026-11-06
- Baidu open-sources ERNIE-5.5-Flash-Base: 280B MoE with 14B active params and 1M context — kimmonismus · 2026-11-06