OpenAI Unveils GPT-5.4-Cyber to Enhance Cybersecurity Defense Capabilities
OpenAI has launched GPT-5.4-Cyber, a variant of its GPT-5.4 model, optimized for cybersecurity applications. This release aims to assist security teams in identifying and addressing vulnerabilities in digital infrastructure. The model is part of OpenAI's Trusted Access for Cyber (TAC) program, which is expanding to include thousands of authenticated defenders and teams. The initiative seeks to democratize access to advanced AI tools while minimizing misuse by bad actors. OpenAI emphasizes the dual-use nature of AI technologies, which can be repurposed for malicious activities. The company aims to provide responsible access to its models, enabling defenders to stay ahead of potential threats while strengthening safeguards against misuse. 
