The existence of a mass gap of 3–5 M _⊙ between the heaviest neutron stars (NSs) and the lightest black holes (BHs), inferred from the BH mass distribution in low-mass X-ray binaries (LMXBs), has been suggested for decades. The gravitational-wave (GW) source GW230529 is most likely a neutron star–black hole (NSBH) merger, with the BH mass falling within this gap. This detection strongly challenges the existence of the gap and has implications for the NSBH population, including a revised BH mass distribution and an updated local merger rate. In this study, we employ POSYDON , a binary population synthesis code that integrates detailed single- and binary-star models, to investigate coalescing NSBH binaries, focusing on the BH mass distribution of the intrinsic NSBH merger population. For typical population models, we find that better matching the observed BH mass distribution requires the use of common-envelope (CE) efficiencies exceeding unity, a rather uncomfortable choice since most energy sources are already included. Alternatively, we find that a two-stage CE prescription calibrated to 1D hydrodynamic simulations has a similar effect. Moreover, motivated by a possible explanation for the firm presence of the mass gap in the LMXB sample, we examine models where the NS birth mass is limited to ≲2 M _⊙ and find excellent agreement with GW observations. Additionally, we present observable distributions of NSBH merger properties, finding good agreement with the measured properties of observed systems and a predicted fraction of NSBH mergers with potential electromagnetic counterparts that ranges from 1% all the way up to 32%.