Context. Meneghetti et al. (2020) (hereafter M20) found that observed cluster member galaxies are more compact than their counterparts in ΛCDM hydrodynamic simulations, as indicated by the difference in their strong gravitational lensing properties. M20 reported that the measured and simulated galaxy-galaxy strong lensing events on small scales are discrepant by one order of magnitude. Among possible explanations for this discrepancy, some studies suggested that simulations with better resolution and implementing different schemes for galaxy formation, compared to the ones used in M20, could bring simulations in better agreement with the observations. Aims. In this paper, we assess the impact of numerical resolution and of the implementation of energy input from AGN feedback models on the inner structure of cluster sub-haloes in hydrodynamic simulations. Methods. We compare several zoom-in re-simulations of a sub-sample of the cluster-sized haloes studied in M20, obtained by varying mass resolution, softening length and AGN energy feedback scheme. We study the impact of these different setups on the subhalo (SH) abundances, their radial distribution, their density and mass profiles and the relation between the maximum circular velocity, which is a proxy for SH compactness. Results. Regardless of the adopted numerical resolution and feedback model, SHs with masses M SH 10 11 h −1 M , the most relevant mass-range for galaxy-galaxy strong lensing, have maximum circular velocities ∼ 30% smaller than those measured from strong lensing observations of Bergamini et al. (2019). We also find that simulations with less effective AGN energy feedback produce massive SHs (M SH 10 11 h −1 M ) with higher maximum circular velocity and that their V max − M SH relation approaches the observed one. However the stellar-mass number count of these objects exceeds the one found in observations and we find that the compactness of these simulated SHs is the result of an extremely over-efficient star formation in their cores, also leading to larger-than-observed SH stellar mass. Conclusions. Regardless of the resolution and galaxy formation model adopted, simulations are unable to simultaneously reproduce the observed stellar masses and compactness (or maximum circular velocities) of cluster galaxies. Thus, the discrepancy between theory and observations that emerged from the analysis of M20 persists. It remains an open question as to whether such a discrepancy reflects limitations of the current implementation of galaxy formation models or the ΛCDM paradigm.