Payward becomes the first crypto company to get access to Claude Mythos 5 through Anthropic's Glasswing program, scanning for code vulnerabilities after 40+ firms demanded AI labs open tools to defenders.
Kraken parent company Payward has joined Anthropic's Project Glasswing and is using Claude Mythos 5 to scan for security vulnerabilities. The company said Monday it's the first crypto firm to gain access to the model, which will review Payward's systems for flaws and send findings to its security teams.
The move follows an open letter last week from more than 40 Bitcoin and crypto companies – including Kraken, Ark Invest, Coinbase, Block and BitGo – that demanded frontier AI labs open trusted-access programs to qualified defenders. Organized by the Bitcoin Policy Institute, the letter argued that developers protecting open-source financial infrastructure need access to frontier models to find and fix vulnerabilities before attackers exploit them.
“Selection gives Payward’s security division early access to the same class of model, sharpening its ability to combat sophisticated software vulnerabilities and protect millions of customers across the globe,” Payward wrote in a statement.
Project Glasswing is Anthropic's cybersecurity initiative, launched in April and expanded in June. It gives vetted organizations access to its most capable cyber models. Partners have uncovered thousands of high- or critical-severity vulnerabilities since the program started.
Payward said any flaws found in third-party open-source software will be reported to the project maintainers. The company added that the approach could help protect its own financial infrastructure and strengthen open-source software used across the crypto industry.
The news comes as AI models show growing utility for both finding and exploiting software vulnerabilities. In April, Mozilla reported that Claude Mythos identified 271 vulnerabilities in Firefox during testing.
“Security has always been an unfair game. An attacker needs to find one flaw. A defender has to find all of them, first, every single day,” Payward Co-CEO Arjun Sethi said in the release. “Frontier AI is the first thing that flips that asymmetry. A model can read every line of code the way an attacker would, at machine scale, so we find the flaw before anyone can build the exploit.”
Anthropic did not immediately respond to a request for comment by Decrypt on Monday.
Drafted by a large language model from the source reporting linked above, then screened by automated publishing checks. It is not read by a journalist before publication. Some articles cite our Alpha Score. Verify prices and figures against the original source. Educational coverage, not personalized advice.