Distribution between well-stirred compartments is the classical paradigm in pharmacokinetics. Also in capillary-issue exchange modeling a barrier-limited approach is mostly adopted. As a consequence of tissue binding, however, drug distribution cannot be regarded as instantaneous even at the cellular level and the distribution process consists of at least two components: transmembrane exchange and cytoplasmic transport. Two concepts have been proposed for the cytoplasmic distribution process of hydrophobic or amphipathic molecules, (i) slowing of diffusion due to instantaneous binding to immobile cellular structures and (ii) slow binding after instantaneous distribution throughout the cytosol. The purpose of this study was to develop a general approach for comparing both models using a stochastic model of intra- and extravascular drug distribution. Criteria for model discrimination are developed using the first three central moments (mean, variance, and skewness) of the cellular residence time and organ transit time distribution, respectively. After matching the models for the relative dispersion the remaining differences in relative skewness are predicted, discussing the relative roles of membrane permeability, cellular binding and cytoplasmic transport. It is shown under which conditions the models are indistinguishable on the basis of venous organ outflow concentration-time curves. The relative dispersion of cellular residence times is introduced as a model-independent measure of cytoplasmic equilibration kinetics, which indicates whether diffusion through the cytoplasm is rate limiting. If differences in outflow curve shapes (their relative skewness) cannot be detected, independent information on binding and/or diffusion kinetics is necessary to avoid model misspecification. The method is applied to previously published hepatic outflow data of enalaprilat, triiodothyronine, and diclofenac. It provides a general framework for the modeling of cellular pharmacokinetics.