OpenAI introduced GPT-5.4-Cyber, a cybersecurity-focused AI model designed for defensive security work. The company said the model will be rolled out on a limited basis to vetted researchers and security vendors due to its advanced capabilities. The launch follows Anthropic’s release of Mythos, which identified thousands of vulnerabilities across software and operating systems. GPT-5.4-Cyber is designed to help detect threats and strengthen cybersecurity defenses. The announcement highlights intensifying competition between AI companies developing specialized models for cybersecurity and vulnerability detection.
OpenAI introduced GPT-5.4-Cyber, a cybersecurity-focused AI model designed for defensive security work. The company said the model will be rolled out on a limited basis to vetted researchers and security vendors due to its advanced capabilities. The launch follows Anthropic’s release of Mythos, which identified thousands of vulnerabilities across software and operating systems. GPT-5.4-Cyber is designed to help detect threats and strengthen cybersecurity defenses. The announcement highlights intensifying competition between AI companies developing specialized models for cybersecurity and vulnerability detection.