OpenAI grants Japan’s megabanks GPT-5.5 for cyberdefense

OpenAI granted Japan’s three largest banks access to its GPT-5.5-Cyber model through a vetted “Trusted Access for Cyber” program, Finance Minister Satsuki Katayama said Friday.reuters
MUFG Bank ↗1.59%, Sumitomo Mitsui Banking ↗2.81%, and Mizuho Bank ↗1.58% are expected to use the model to probe their own systems for security gaps, according to Nikkei.nikkei
The deal, brokered after U.S.-Japan government discussions, also opens a path for Japanese institutions to access Anthropic’s Claude Mythos model following Trump administration approval, Katayama said.morningstar

OpenAI Grants Japan’s Megabanks Access to GPT-5.5 for Cyberdefense

Japanese Finance Minister Satsuki Katayama announced on Friday that OpenAI has granted select Japanese financial institutions access to its GPT-5.5 model to bolster defenses against cyberattacks, following a meeting with OpenAI Chief Strategy Officer Jason Kwon in Tokyo.reuters

Without naming specific institutions, Katayama described the access as “a big step forward in strengthening Japanese financial institutions’ ability to defend against cyberattacks,” according to Reuters.straitstimes

Japan’s Biggest Banks Expected to Benefit

The Nikkei newspaper reported on May 28 that Japan’s three largest banks — MUFG Bank, Sumitomo Mitsui Banking Corp, and Mizuho Bank — are expected to gain access to the model. The banks have not independently confirmed the arrangement.nikkei

The model, referred to as GPT-5.5-Cyber, is being delivered through OpenAI’s “Trusted Access for Cyber” programme, a framework designed to ensure only verified defenders can use the tool. According to Dow Jones reporting via Morningstar, the model could allow Japan to better scrutinize systemic risks, identify gaps in information security, and implement countermeasures.thenextweb

Broader Context

The decision followed discussions between the Japanese and U.S. governments and mirrors a similar agreement previously reached with European companies. Katayama also said Japan’s government and financial institutions were expected to gain access to Anthropic’s Claude Mythos model to further enhance their defenses.morningstar

The Nikkei reported that GPT-5.5 is believed to be comparable in capability to Anthropic’s Claude Mythos, which is currently limited to roughly 50 companies worldwide. Some Japanese financial firms will also gain access to Mythos following approval from the Trump administration, Katayama said.straitstimes

The move reflects a growing recognition among regulators that the same frontier AI systems enabling more sophisticated cyberattacks must also be deployed defensively. “With frontier AI now recognized as a threat, this step could serve as a welcome catalyst to strengthen cybersecurity within Japan’s financial sector,” Katayama told reporters.morningstar