Claude Mythos Preview finds new cryptanalytic attacks on HAWK and AES
新智元 · wechat · 2026-07-29
Anthropic’s Claude Mythos Preview reportedly found new attacks on two cryptographic targets: HAWK, a NIST post-quantum signature candidate, and simplified AES-128.
- On HAWK, the model identified a previously unused symmetry in the lattice structure and cut the estimated security from about 2^64 operations to 2^38.
- On simplified AES-128, it devised a new attack called “Möbius Bridge”, improving the best known approach by 200–800x.
- The article says both results cost roughly $100k in API usage, and human researchers needed weeks to verify them.
- It also notes a further claim: the model found a practical attack on 13-round LEA, reducing a previously infeasible computation to about an hour on a normal computer.
The key theme is that AI is moving from implementation bugs to math-level cryptanalysis, shifting the bottleneck from discovery to verification.
Related event: Anthropic's Claude Finds Weaknesses in HAWK and Reduced-Round AES(39 posts)→
More from Safety
- Dev builds MCP layer so Claude Code can work on servers without ever holding SSH keys — NETMESAFE · 2026-10-16
- 'Human in the loop' called one of AI governance's biggest lies: presence isn't authority — glenbeer · 2026-10-16
- Tencent eval agents wandered off-task, India CERT warns: rogue agent stories keep coming — ur_piyo_a_hoe · 2026-10-16
- AI containment bingo: 7 squares crossed off in a single week — AIGODSEND · 2026-10-16
- OpenAI cyber eval agents escaped to the internet and compromised Hugging Face infrastructure — nathanbenaich · 2026-10-16
- UK Dev Asks How the Online Safety Act Applies to a Venice AI-Style Aggregator — Every-Price-4504 · 2026-10-16