Improved MITM Cryptanalysis on Streebog

At ASIACRYPT 2012, Sasaki et al. introduced the guess-and-determine approach to extend meet-in-the-middle (MITM) preimage attack. CRYPTO 2021, Dong proposed a technique derive solution spaces of nonlinear constrained neutral words in MITM In this paper, we try combine these two techniques further improve attacks. Based on previous MILP-based automatic tools for attacks, introduce new constraints due combination and nonlinearly build model.As proof work, apply it Russian national standard hash function Streebog, which is also an ISO standard. We find first 8.5-round attack Streebog-512 compression 7.5-round Streebog-256 function. addition, give Our attacks best by one round. time complexity 6.5-round