Around The IT World

Professional Tech News from 23+ Languages

🇭🇰 Chinese, 5 November 2026, AI-powered security

OpenAI Releases GPT-6 Cyber: AI Capabilities in Cybersecurity Leap Forward

OpenAI's GPT-6 Cyber AI model has significantly improved its ability to detect and exploit vulnerabilities.

OpenAI’s GPT-6 Cyber AI model has shown a 51% improvement in its vulnerability exploitation benchmark score compared to its predecessor. It has assisted researchers in discovering 81 high-risk vulnerabilities in Firefox and has developed exploit programs for 948 Linux kernel vulnerabilities, a five-fold increase from its predecessor. The model’s out-of-bound attack rate has decreased to 0.07% and its security capabilities have been upgraded to a ‘critical’ level. OpenAI has restricted access to the model to only verified clients due to concerns over potential misuse of its enhanced capabilities.

Back to the latest news