Packed bed reactors are common reactor types in chemical industries. In case of reactions with large heat effects, bundles of slender tubes are used. For such slender tubes, the column-to-particle diameter ratio, N, is small, which can give rise to significant flow channelling. This research uses two particle-scale numerical approaches, namely particle-resolved computational fluid dynamics (PR-CFD) and pore network model (PNM), to investigate the hydrodynamics of three packed beds containing random packings of spherical particles with 4.2⩽N⩽7.0. Global parameters of the PNM, such as shape and constriction factors, are optimized using the PR-CFD results. A comparison of computed PNM and PR-CFD results shows that the PNM captures local variations in the bed well. The low computational cost of PNM, as well as its ability to provide locally resolved data, makes the PNM a promising approach for the pore-scale modelling of slender packed bed reactors.