Japan’s three megabanks—Mitsubishi UFJ Financial Group (MUFG), Sumitomo Mitsui Financial Group, and Mizuho Financial Group—are set to gain access to Anthropic’s advanced AI model, Mythos, by the end of May 2026 [1, 2, 3, 4].
The banks likely received confirmation of this access from U.S. Treasury Secretary Scott Bessent during his meeting with Japanese Finance Minister Satsuki Katayama in Tokyo on May 12, 2026 [1, 3, 5]. Mythos has so far been available to a small number of U.S. banks and organizations, with planned expansion to major banks in Japan, Europe, and the UK [5, 4].
Mythos is designed to detect cybersecurity vulnerabilities that are challenging for humans to identify, a capability that has raised concerns among financial institutions and regulators due to its potential impact on financial system security [5, 6, 4].
In response, Japanese Finance Minister Katayama announced that Japan would establish a public-private working group this week to address the cybersecurity risks posed by Mythos. The group’s first meeting took place on May 14, 2026, in Tokyo. Katayama said, “Japan would establish a public-private working group this week to address cybersecurity risks to the Japanese financial system posed by Mythos” [5].
The three megabanks declined to comment on the decision to gain access to Mythos, and Anthropic did not respond to requests for comment [5].
The rollout will mark one of the first major uses of Mythos outside the United States. Access will officially begin by the end of May 2026 [1, 2, 3, 4].
