Japanese banks to get Anthropic’s vulnerability-hunting AI

MUFG, Mizuho, and SMFG would be the first Japanese institutions added to Anthropic’s restricted Project Glasswing rollout, a source familiar with the matter told Reuters

Japan’s three megabanks are set to gain access to Claude Mythos, Anthropic’s vulnerability-hunting AI model, within roughly two weeks, a source familiar with the matter told Reuters on Tuesday.

It would be the first time a Japanese company has been granted entry to the restricted preview, which has so far been confined to Anthropic’s American and a handful of European partners.

Mitsubishi UFJ Financial Group, Mizuho Financial Group, and Sumitomo Mitsui Financial Group were informed of the move during meetings in Tokyo this week with US Treasury Secretary Scott Bessent. The three lenders are expected to be onboarded by the end of May.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

Mythos has been treated by regulators and chief executives as a category-shifting event since Anthropic disclosed its existence earlier this month.

The model has discovered thousands of previously unknown zero-day vulnerabilities across every major operating system and every major web browser, and in internal testing it wrote working exploits, including chains that escape both renderer and operating-system sandboxes in a browser.

Mozilla last week shipped Firefox 150 with fixes for 271 vulnerabilities found by Mythos in a single evaluation pass.

Anthropic has not released the model publicly. Instead, it has run a controlled rollout under what it calls Project Glasswing, with 12 named launch partners, including AWS, Apple, Cisco, Google, JPMorganChase, Microsoft, Nvidia, and Palo Alto Networks, and around 40 further institutions granted access on a case-by-case basis.

Japan’s inclusion comes weeks after the Fed and US Treasury convened American bank chief executives on the same cyber-risk briefing, and after UK regulators committed to briefing major British banks within days.

Tokyo is moving in parallel. Finance Minister Satsuki Katayama announced the formation of a 36-entity public-private working group on Mythos-class risks, comprising the country’s major banks, the Bank of Japan, and the Japanese units of Anthropic and OpenAI.

The group is chaired by Mizuho’s chief information security officer and is charged with identifying exposures, implementing defensive measures, and drafting contingency plans for what would amount to a co-ordinated patching push across the Japanese financial system.

For the three banks involved, the immediate question is operational. Mythos under Glasswing terms is delivered with restrictions on output disclosure, with the model used to find vulnerabilities in a partner’s own systems and to draft remediation, not to publish exploits.

The Mozilla case offers a template: 271 vulnerabilities patched in a single Firefox release after a Mythos sweep, with the model’s findings handed back to Mozilla engineers under non-disclosure rather than published.

The geopolitical layer is unusually visible. Bessent’s role in conveying the access decision in Tokyo aligns Mythos rollout with US Treasury statecraft rather than with Anthropic’s commercial channel, an arrangement that has drawn complaints from European capitals.

Eurozone finance ministers raised the issue at an Ecofin meeting last week, where no EU government had access to the model while the White House was reported to be blocking further expansion of the partner list.

Industry views on Mythos remain split. Some cybersecurity researchers have argued that the vulnerabilities Mythos surfaced are reachable through clever orchestration of public models, and that the bigger story is the rate of improvement of frontier AI in offensive cyber, not Mythos itself.

Others, including Anthropic chief executive Dario Amodei, have described the moment as a “cyber moment of danger” that justifies the access controls.

Anthropic and the three Japanese banks did not immediately respond to requests for comment, according to the Reuters source’s account.