This paper reports the methodology used for the neutronic modelling of the CROCUS reactor and discusses the challenges encountered during the process. Full-core steady-state neutronics solutions were computed with the PARCS code. The Serpent Monte Carlo code was used for few-group constant generation. The full-core Serpent model of the reactor was also used as reference for the comparison against PARCS results. The comparison between Serpent and PARCS solutions was successful, achieving good level of agreement for eigenvalue (418 pcm dierence) and control rod reactivity worth (1 pcm dierence). In terms of radial neutron flux profiles, dierences in the inner fuel region were within 5% and 1% for the thermal and fast fluxes respectively. However, in the outer fuel lattice region, dierences were considerably higher due to the mismatch between PARCS nodes and heterogeneous fuel pins. Also, PARCS post-processing for intranodal reconstruction proved to be an eective way to observe heterogeneities within nodes, which cannot be otherwise captured by PARCS solution. Some of the modelling challenges were overcome with the use of transport-corrected diusion coecients and the implementation of albedo boundary conditions. A parametric analysis reflected the importance of the transport correction of diusion coecients for producing good eigenvalues in reactor cores with large neutron leakage.