Shock-induced energy localization is a crucial mechanism for determining shock sensitivity of energetic materials (EMs). Hotspots, i.e., localized areas of elevated temperature, arise when shocks interact with defects (cracks, pores, and interfaces) in the EM microstructure. The ignition and growth of hotspots in a shocked energetic material contribute to rapid chemical reactions that can couple with the passing shock wave, potentially leading to a self-sustained detonation wave. Predictive models for shock-to-detonation transition must correctly capture hotspot dynamics, which demands high-fidelity material models for meso-scale calculations. In this work, we deploy atomistics-guided material models for the energetic crystal RDX (1,3,5-trinitro-1,3,5-triazinane) and perform tandem continuum and all-atom molecular dynamics (MD) simulations. The computational setup for the continuum and MD simulations are nearly identical. The material models used for the calculations are derived from MD data, particularly the equations of state, rate-dependent Johnson–Cook strength model, and pressure-dependent shear modulus and melting temperature. We show that a modified Johnson–Cook model that accounts for shear-induced localization at the pore surface is necessary to represent well—relative to MD as the ground truth—the inelastic response of the crystal under a range of shock conditions. A head-to-head comparison of continuum and atomistic calculations across several metrics of pore collapse and energy deposition demonstrates that the continuum calculations are in good overall agreement with MD. Therefore, this work provides improved RDX material models to perform physically accurate meso-scale simulations, to enhance understanding of hot spot formation, and to use meso-scale hot spot data to inform macro-scale shock simulations.