HODL Hard, Math Harder: Claude Mythos Pecks HAWK Before It Leaves the Nest 🪨
Back to feed

HODL Hard, Math Harder: Claude Mythos Pecks HAWK Before It Leaves the Nest 🪨

Anthropic said an unreleased version of its most powerful AI model, Claude Mythos Preview, identified two previously unknown attacks on cryptographic algorithms, one of them against a scheme currently competing to become a U.S. federal standard. The target, HAWK, is a digital signature system designed to survive future quantum computers and to prove a transaction came from a user without ever exposing their private key. The non-regulatory federal agency and lab NIST moved HAWK into the third round of its post-quantum signature competition in May, where it remains the last lattice-based candidate standing.

Claude found a symmetry buried in HAWK's math that no human had thought to use. For the smallest configuration, the cost of recovering a secret key fell from 2^64 operations to 2^38, roughly 67 million times less work. Patching the vulnerability means roughly doubling HAWK's key sizes. "Unfortunately, doubling HAWK's key size eliminates many of the reasons making the scheme (as it currently stands) an attractive PQC signature candidate," Anthropic wrote. The trade carries real weight on blockchains, where signature size is block space and block space is fees; any chain shopping for a quantum-resistant replacement is partly choosing on bytes per signature, and compact keys plus fast signing were HAWK's entire pitch.

The second attack targets AES, the cipher scrambling HTTPS traffic, encrypted drives, and exchange backends. Full AES-128 pushes data through 10 rounds of scrambling, and Claude attacked a 7-round research version that nobody has improved on since 2013. Researchers barred the model from all five established families of AES cryptanalysis and asked it to invent a sixth, closing the brief with a line about how the first differential attack did not beat anything, it invented the game. The model also inherited working notes from earlier agent runs that had already burned through roughly 200 failed attack variants. It refused anyway. "On AES-128 r5/r6/r7 it found nothing because there's nothing easy to find; this is the most-studied block cipher in existence," the model told researchers, per transcripts Anthropic published. Anthropic disclosed both results to the algorithms' authors and to U.S. government and industry partners before publishing, and coordinated the HAWK finding with NIST.

For crypto holders, HAWK has never been deployed anywhere, and $BTC still runs on ECDSA, the pre-quantum signature scheme that candidates like HAWK are eventually meant to replace, so nothing in production is broken. Anthropic published its findings on July 28, 2026, and described the research as a step toward using the same tool to defend as well as attack.

Share:
Publishercryptonewsroom.xyz
Published—
CategorySecurity

Disclaimer: This content is for information and entertainment purposes only. It does not constitute financial, investment, legal, or tax advice. Always do your own research and consult with qualified professionals before making any financial decisions.

See our Terms of Service, Privacy Policy, and Editorial Policy.