OpenAI Launches GPT-5.4 with Reverse Engineering, Vulnerability and Malware Analysis Features

OpenAI announced GPT-5.4-Cyber, a GPT-5.4 variant fine-tuned to support vetted security professionals with advanced defensive workflows (including binary reverse engineering, vulnerability scanning, and malware analysis), alongside a scaled Trusted Access for Cyber (TAC) program that expands tiered access for verified individuals and teams; the report also highlights Codex Security’s role in fixing vulnerabilities and emphasizes safeguards and dual-use risks but does not report any incident or active exploitation.