According to monitoring by Dongcha Beating, OpenAI has announced the first cohort of partner organizations for its 'Trusted Access for Cyber' program. Among the 14 institutions, nearly half are from Wall Street: Bank of America, BlackRock, BNY Mellon, Citigroup, Goldman Sachs, JPMorgan Chase, and Morgan Stanley, along with cybersecurity companies CrowdStrike, Zscaler, SpecterOps, iVerify, as well as Cisco, NVIDIA, and Oracle. These institutions will gain access to GPT-5.4-Cyber for securing their respective digital infrastructures. The concentration of financial institutions is not surprising, as banks are among the most targeted industries for cyberattacks globally, with large security teams, substantial budgets, and a rapid adoption of new tools, making them highly motivated to engage early in AI security tools for influence. Leigh-Ann Russell, Chief Information Officer of BNY Mellon, stated in a release that the bank is closely collaborating with 'the forefront forces driving these efforts.' Additionally, OpenAI announced a $10 million API grant through its 'Cybersecurity Grant Program' aimed at open-source security and vulnerability research teams. The first recipients of the funding include software supply chain security company Socket and Semgrep, as well as vulnerability research organizations Calif and Trail of Bits. OpenAI mentioned in a blog post that not every organization has a 24/7 security team capable of responding promptly when vulnerabilities are disclosed on 'Friday evenings,' which is precisely the issue the grant program aims to address. OpenAI also provided access to GPT-5.4-Cyber to the U.S. Center for AI Standards and Innovation (CAISI) and the UK AI Security Institute (UK AISI) for both organizations to assess the model's cybersecurity capabilities and protective measures.