An AI model just found real math holes in two cryptosystems that experts had reviewed for years. Not a jailbreak, not a demo — actual novel attacks. Anthropic’s Frontier Red Team published it on July 28.
What Claude Mythos actually did
Claude Mythos is Anthropic’s unreleased frontier model, and here it worked as a semi-autonomous research agent. On HAWK, a NIST post-quantum signature candidate, it improved the best attack from 2^64 down to 2^38 operations — gutting the smallest key’s effective strength and forcing researchers to double key sizes. That took ~60 hours and roughly $100K in API spend. On AES, it invented a fingerprinting trick called “Möbius Bridge” that speeds up a meet-in-the-middle attack on 7-round AES-128 by 200–800x, running almost fully autonomous for three days.
Why it matters
Nothing in production is broken — HAWK isn’t deployed and the AES attack only hits a reduced-round version. The point is the capability. A model steered by one human doing project management, not math, out-cranked years of expert review. This is Claude Mythos delivered through the Anthropic API, positioned as a research collaborator for the hardest formal problems, not a chatbot.
You Might Also Like
- Anthropic Claude Security Goes Public Beta After 500 Cves Found in Closed Preview
- Anthropic Just Launched Code Review in Claude Code and 54 of prs now get Real Feedback
- Claude Replay Turns Your Anthropic Claude Code Sessions Into Shareable Video Like Replays
- 13b Into Openai yet Microsoft Copilot Cowork Runs on Anthropic Claude
- Anthropic Gives Claude Computer use on mac and you can Control it From Your Phone
Top AI Product 
Leave a comment