Payward, the parent company of the cryptocurrency platform Kraken, has joined Project Glasswing, Anthropic's initiative that allows selected companies to use restricted frontier artificial intelligence models to secure critical software and infrastructure against cyber threats. The Wyoming-based company announced on Monday that it plans to deploy Claude Mythos 5 in its environments in the coming weeks, with findings that will feed into its existing security programme.
Project Glasswing already includes technology giants such as Amazon Web Services, Apple, Google and Microsoft, as well as financial institutions like JPMorganChase. Payward's move comes after the US government made Mythos 5 available to American organisations that operate and defend critical infrastructure. Anthropic launched Project Glasswing in April 2026 and has since expanded the initiative to organisations in technology and finance.
Payward also plans to share vulnerabilities found in third-party open-source software with the relevant project maintainers. Artificial intelligence is emerging as a growing threat to cryptocurrency companies, providing attackers with tools to find vulnerabilities faster, automate attacks and make phishing campaigns and social engineering more convincing.
Payward argued that cryptocurrency platforms face security challenges similar to other critical financial infrastructure. Exchanges, custody systems and settlement rails operate continuously and can be lucrative targets for attackers. "Security has always been an unfair game. An attacker needs to find one flaw. A defender has to find them all, first, every day," said Arjun Sethi, co-president of Payward, in the statement.
Claude Mythos 5 is Anthropic's most advanced model for defensive cybersecurity, designed to analyse code at scale, identify vulnerabilities and help developers repair them. Anthropic has indicated that it plans to expand access to Mythos-like cybersecurity capabilities as it develops safeguards for broader deployment.