Get Zlio on your home screen
One tap, full screen, like an app.
A frontier lab shipping a purpose-built cyber model raises the bar for AI-driven attack and defence.
OpenAI released GPT-6 Cyber, a model for defensive cybersecurity work that it says is less likely to be misused for hacking, with a 'Critical' rating from a cyber capability evaluator. It also described a pipeline that found and patched 948 Linux kernel privilege-escalation flaws.
Read it at OpenAI → 💬 356 on HN