MiniMax says internal model cracked multiple NIST candidates with no known attacks
MiniMax 稀宇科技 · wechat · 2026-11-22
MiniMax reports that its latest internal model performed systematic automated cryptanalysis, achieving several firsts:
- NIST competition candidates: On CryptanalysisBench Tier2 (142 schemes with no known practical attacks), the model produced the first practical collision attack on Waterfall-256 (0.045s), full-round collisions on Blender-256 (0.36–2.3s), and key-recovery breaks on post-quantum candidates LOCKER and ROLLO-I. It also independently found 10+ previously unknown severe flaws in official reference code of DRS, RQC-128, LAC-128, KCL, LEDAkem, ISAP and others.
- Academic frontier schemes: New forgery attack on CAESAR winner COLM (2 encryption queries), 0.06s private-key recovery on ASIACRYPT 2025's SBC-VOLE, and pushing practical key recovery on NSA's ARADI from round 5 to round 8.
- Classical hard problems: Recovered keys for 94 TI McEliece challenge instances (55 first-ever solves), 6,180–43,400x faster than the CRYPTO 2026 record; HAWK lattice solving 690x more efficient than Anthropic's approach.
MiniMax says current standards (AES, RSA, SHA-256) are unaffected, findings were responsibly disclosed to authors, and the technical report and code are open-sourced, with test access available for researchers.
More from Models
- MiniMax's internal model claims first practical breaks on NIST candidates, up to 43,400x faster than Anthropic solvers — Dan_Jeffries1 · 2026-11-22
- Google testing Inbox mode for Antigravity to shift users into approving AI outcomes — testingcatalog · 2026-11-22
- Can you resell ChatGPT Pro's $200 monthly API credits in a commercial SaaS? — infolily · 2026-11-22
- Insider teases: top AI accounts are early-testing an "absolute monster" model — iruletheworldmo · 2026-11-22
- Inkling 2 mini's 1M-context KV cache dissected: only 8 of 48 layer visits keep full cache — teortaxesTex · 2026-11-22
- GPT-6.2 is hard to steer via system prompt — OpenAI says use it directly in Codex — pvncher · 2026-11-22