MiniMax says internal research model found unknown crypto flaws, achieved first practical breaks on NIST candidates

StefanoGogioso · x · 2026-11-22

MiniMax Research used an internal-only model (MIM) for automated cryptanalysis of modern cryptographic schemes, uncovering previously unknown structural flaws and turning theoretical attacks into practical breaks.

The team frames this as cryptanalysis targeting algorithm design itself rather than implementation bugs, with implications for securing communications, finance, and critical infrastructure.

Related event: MiniMax internal model breaks NIST candidate algorithms up to 1000x faster(2 posts)→

Original post →

More from Research

Research channel →