OpenAI Extends GPT-5.4-Cyber Access to Trusted Organizations Worldwide

OpenAI announced the expansion of its "Trusted Access for Cyber" program to grant security organizations worldwide access to the GPT-5.4-Cyber model, backed by $10 million in API credits for defenders. The announcement highlights initial grant recipients (e.g., Socket, Semgrep, Calif, Trail of Bits), partnerships with major financial and tech firms, and involvement of U.S. and U.K. AI security standards bodies, emphasizing defensive research, safety guardrails, and broader defender access.