An adequate representation of aqueous solvent is a fundamental problem in the field of macromolecular simulations. To assess the ability of different solvent models to reproduce experimental data, we performed a series of molecular dynamics simulations on a small peptide, each utilizing a different model of solvent. The generalized Born (GB) model, the analytical continuum electrostatics (ACE) potential, the effective energy function-1 (EEF1), the solvent accessible surface area (SASA) model, and the standard TIP3P model of explicit solvent were evaluated in this study. For each solvent model, the potential of mean force (pmf) for folding a 12-residue peptide from the fully extended state to a compact state, where the two ends of the peptide are in close proximity, was computed. These data were compared to experimental results on the peptide's end-to-end distance distribution obtained with fluorescent resonance energy transfer (FRET) experiments. For each solvent model, the FRET efficiency was computed from the corresponding pmf and compared to the experimental result. The value obtained from the simulation with explicit solvent was in excellent agreement with experiment. By contrast, all simulations that employed an implicit solvent model yielded values for the FRET efficiency that significantly deviated from the experimental value. An analysis of the energetic contributions to the pmf suggests potential etiologies for this marked discrepancy from experiment.