Temperature-based replica-exchange molecular dynamics (REMD), in which multiple simultaneous simulations, or replicas, are run at a range of temperatures, has become increasingly popular for exploring the energy landscape of biomolecular systems. The practical application of REMD toward systems of biomedical interest is often limited by the rapidly increasing number of replicas needed to model systems of larger size. Continuum solvent models, which replace the explicit modeling of solvent molecules with a mean-field approximation of solvation, decrease system size and correspondingly, the number of replicas, but can sometimes produce distortions of the free energy landscape. We present a hybrid implicit/explicit solvent REMD method in CHARMM in which replicas run in a purely explicit solvent regime while exchanges are implemented with a high-density GBMV2 implicit solvation model. Such a hybrid approach may be able to decrease the number of replicas needed to model larger systems while maintaining the accuracy of explicit solvent simulations. Toward that end, we run REMD using implicit solvent, explicit solvent, and our hybrid method, on three model systems: alanine dipeptide, a zwitterionic tetra-peptide, and a 10-residue β-hairpin peptide. We compare free energy landscape in each system derived from a variety of metrics including dihedral torsion angles, salt-bridge distance, and folding stability, and perform clustering to characterize the resulting structural ensembles. Our results identify discrepancies in the free-energy landscape between implicit and explicit solvent and evaluate the capability of the hybrid approach to decrease the number of replicas needed for REMD while reproducing the energy landscape of explicit solvent simulations.