OpenAI has announced the deployment of its GPT-5.5-Cyber model to Japan's largest banks, including MUFG Bank, Sumitomo Mitsui Banking Corp, and Mizuho Bank, to bolster their cybersecurity defenses. This initiative, revealed on May 21, is part of a broader collaboration with the Japanese government and select corporations, aimed at enhancing protection against cyber threats. The deployment comes as Anthropic's Claude Mythos, capable of autonomously discovering zero-day vulnerabilities, becomes available to these banks by the end of May 2026. The collaboration involves Japan's Finance Minister Satsuki Katayama and US Treasury Secretary Scott Bessent, highlighting the international dimension of this cybersecurity effort. OpenAI's "Trusted Access for Cyber" program ensures that only verified entities can utilize these advanced AI tools, addressing concerns about dual-use capabilities. This move signals a shift towards controlled distribution of AI technologies, potentially setting a precedent for future regulatory frameworks in AI cybersecurity.