OpenAI has given Japan's three biggest banks—MUFG Bank, Sumitomo Mitsui Banking Corp, and Mizuho Bank—access to its GPT-5.5 model aimed at enhancing cybersecurity defenses, Japanese Finance Minister Satsuki Katayama announced on May 29, 2026, after meeting OpenAI’s chief strategy officer Jason Kwon in Tokyo [1, 2].
The GPT-5.5 model is highly capable in advanced coding, a strength that has unfortunately also enabled hackers to identify and exploit cybersecurity weaknesses more efficiently [1, 2]. Katayama called the access arrangement “a big step forward in strengthening Japanese financial institutions’ ability to defend against cyberattacks,” highlighting the significance of this collaboration [1, 2].
The Nikkei newspaper reported a day earlier that Japan’s three largest financial institutions were expected to receive the model, marking an important development in AI-driven cybersecurity for the nation’s banking sector [1].
Alongside OpenAI’s model, Japan’s government and financial sector are expected to gain access to Anthropic’s Claude Mythos AI cybersecurity tool to further defend against threats [1, 2]. In May 2026, Japan set up a public-private working group specifically to address the risks posed by AI models like Mythos to the country’s financial systems [1, 2].
The decision by OpenAI followed negotiations between the Japanese and U.S. governments, coming after similar access had been granted to some European companies [1, 2]. This represents an expansion of advanced AI tools offered to key international financial actors amid rising cybersecurity concerns.
Requests for comment from MUFG Bank and Mizuho Bank were declined; Sumitomo Mitsui Banking Corp was not immediately available for comment [1, 2].
The next concrete step under this initiative is the ongoing collaboration and deployment of both OpenAI’s GPT-5.5 and Anthropic’s Claude Mythos within Japan’s financial institutions to strengthen cybersecurity measures in the months ahead.
