The contribution of differential absorption-distribution pharmacokinetics to drug activity can be partially determined by comparing simultaneous estimates of drug serum level with pharmacodynamic effects. In the present paper we have contrasted the effects of clinically equipotent doses of lorazepam, alprazolam, and diazepam on the performance of tracking and digit symbol substitution tasks. Eight young males were tested for 12 h after ingesting the drug. The three benzodiazepines and placebo were administered to each subject according to a balanced double-blind Latin square design. A model is presented that describes the relationship between drug concentration and the degree of impairment across time after the final peak effect. Exponential rate constants were determined for each drug using a Marquardt nonlinear fit of the pooled data. Basically, the constants relate offset serum drug values to the impairment curves at a time when serum-brain equilibrium is assumed to have occurred. The values indicate markedly rapid acute tolerance for alprazolam and diazepam but relatively little acute tolerance for lorazepam. Whether these constants reflect adaptation or differential association-dissociation receptor rate constants cannot be determined, but they do highlight the need to consider receptor kinetics as an important factor in benzodiazepine pharmacodynamics.