OpenAI Grants Japanese Banks Access to GPT-5.5-Cyber Model

OpenAI has granted Japanese financial institutions access to its GPT-5.5-Cyber model to help strengthen defenses against increasingly sophisticated cyberattacks [1, 4]. The initiative, confirmed by Japanese Finance Minister Satsuki Katayama, follows high-level negotiations between the Japanese and U.S. governments [1, 4].

Key takeaways

• Japan’s three largest banks—MUFG Bank, Sumitomo Mitsui Banking Corp, and Mizuho Bank—are expected to receive access to the model [1, 4].
• The access is provided through OpenAI’s "Trusted Access for Cyber" program, which restricts the most capable AI tools to verified defenders [3, 4].
• The collaboration aims to counter risks posed by new AI models capable of identifying and exploiting system vulnerabilities at a high level [1, 3].
• Japan has established a public-private working group to address cybersecurity risks to the financial system posed by frontier AI models [1, 4].

Strengthening Financial Defenses

The integration of GPT-5.5-Cyber into Japan’s financial sector is designed to provide banks with advanced tools to identify and patch vulnerabilities before they can be exploited by malicious actors [1, 3]. As AI models become more adept at coding, they have granted hackers an unprecedented ability to devise complex cyberattacks, creating an urgent need for defensive systems that operate at a similar technical level [1, 2]. By providing early access to these models, OpenAI aims to empower trusted partners to better protect their infrastructure [1, 3].

This move follows a similar rollout in Europe, where OpenAI granted access to companies including Deutsche Telekom and BBVA [3]. In Japan, the arrangement is viewed as a strategic government-to-government understanding, with Finance Minister Satsuki Katayama and U.S. Treasury Secretary Scott Bessent directly involved in the discussions [4]. Beyond OpenAI, Japanese institutions are also expected to gain access to Anthropic’s Claude Mythos model, further expanding the defensive capabilities available to the country's financial sector [1, 4].

Why it matters

The deployment of frontier AI models as defensive tools marks a shift in how nations treat artificial intelligence, moving from consumer-facing applications to critical national infrastructure [4]. While this program provides major institutions with powerful security resources, it also creates a potential two-tier security landscape [4]. By concentrating advanced defensive AI within a handful of large, vetted banks, smaller financial entities and fintech startups may remain more exposed to the widening gap in cybersecurity capabilities [4]. As the financial system navigates these new threats, the coming months will determine whether this strategy successfully stabilizes the broader economy or primarily benefits the largest market participants [4].