OpenAI has expanded its Trusted Access for Cyber (TAC) program, introducing various tiers for users authenticated as cybersecurity defenders, enabling them to use the new GPT‑5.4‑Cyber model. This iteration is specifically fine-tuned for enhanced cyber capabilities, allowing users to conduct advanced security tasks, such as binary reverse engineering for malware analysis, with fewer restrictions. The TAC initiative emphasizes democratized access, leveraging identity verification processes, and builds upon previous model safeguards that began with GPT-5.2. OpenAI’s ongoing cybersecurity strategy focuses on providing broad access to these capabilities while ensuring safety and protection against misuse through iterative deployment and investments in ecosystem resilience.

OpenAI: OpenAI is an AI research and deployment company focused on developing advanced large language models and tools tailored for specific use cases including cybersecurity defense. In this announcement, OpenAI is expanding its Trusted Access for Cyber program and introducing GPT-5.4-Cyber, a fine-tuned variant of GPT-5.4 designed for cyber-permissive capabilities like binary reverse engineering to aid verified defenders.
Codex Security: Codex Security is an OpenAI system that automatically monitors codebases, validates security issues, and proposes fixes to enhance software vulnerability management. Featured in this update, it represents OpenAI’s ongoing efforts to integrate AI into developer workflows for continuous security improvements as part of their cyber defense strategy.
Trusted Access for Cyber: Trusted Access for Cyber (TAC) is OpenAI’s verification program that provides cybersecurity professionals and teams with access to AI models featuring reduced safeguards for legitimate defensive tasks. In this expansion, TAC introduces new tiers culminating in access to GPT-5.4-Cyber for the most verified users, emphasizing automated identity checks and objective criteria.
Cybersecurity Grant Program: The Cybersecurity Grant Program is OpenAI’s initiative to fund and support cybersecurity researchers and defenders in addressing AI-related security challenges. It forms one of the foundational elements of OpenAI’s long-term strategy outlined in this post, alongside model safeguards and access programs.

`json
{
“TAC Access”: “TAC enables verification for individuals and enterprises to gain reduced-friction access to cyber-focused model capabilities, allowing them to engage in legitimate cybersecurity work with fewer restrictions.”,
“Model Safeguards”: “OpenAI incorporates cyber-specific safeguards with models starting from GPT-5.2 and continues to enhance these safeguards through subsequent versions, recognizing GPT-5.4 for its cyber capabilities under the Preparedness Framework.”,
“Defense Principles”: “OpenAI’s cybersecurity approach is guided by principles of democratized access through verification, iterative deployment for continuous learning and improvement, and strengthening the ecosystem through dedicated tools and grants.”
}
`