- Who
- MiniMax
- What
- reported automated cryptanalysis results using MiniMax Cyber
- Quote
Our research team used MiniMax Cyber to conduct automated cryptanalysis across modern cryptographic schemes.
What happened
MiniMax says its research team used MiniMax Cyber to turn theoretical cryptanalysis into practical attacks on NIST candidates, peer-reviewed schemes, and public cryptography challenges. Reported results include breaking full-round collision resistance in Waterfall-256 and Blender-256, recovering keys from several constructions, and solving 55 previously unsolved McEliece challenge instances. The work focuses on academic designs, past competition candidates, and public challenge parameters; MiniMax says it does not affect deployed cryptographic standards.
Why it matters
MiniMax Cyber turned cryptanalysis into executable attacks against academic schemes and public challenge instances, while MiniMax says the work does not affect deployed cryptographic standards and still requires human review of security margins.
2 reports
The story so far
MiniMax said its research team used MiniMax Cyber and its internal research model, MIM, to conduct automated cryptanalysis and turn theoretical attacks into practical ones against NIST competition candidates, peer-reviewed schemes, and public cryptography challenges. According to MiniMax, the system broke full-round collision resistance in Waterfall-256 and Blender-256, recovered keys from LOCKER and ROLLO-I, forged a COLM ciphertext with two encryption queries, and solved 55 previously unsolved McEliece challenge instances. MiniMax said the work focuses on academic designs, past competition candidates, and public challenge parameters, and does not affect deployed cryptographic standards. The company also released curated and validated reasoning trajectories from the experiments as technical reports and proof-of-concept code at https://github.com/MiniMax-AI/Research, and said qualified researchers and developers can apply for testing access by sending a brief introduction and research background summary to [email protected].
