Animal models of human drug abuse measuring (i) cocaine-seeking behavior maintained by primary or secondary reinforcers, (ii) the time to extinction of cocaine-seeking behavior, and (iii) following extinction, reinstatement of operant responding resulting in presentation of a cocaine-associated stimulus were developed in rats. Animals were trained to respond on a multiple schedule of food and intravenous cocaine reinforcement during which either food or cocaine paired with auditory or visual stimuli were available during four alternating 30-min schedule components. At the beginning of each experimental session, a 10-min component (stimulus component) was introduced during which the stimuli associated with the primary reinforcers were made available response-contingently. Subsequent non-contingent presentation of food or cocaine at the beginning of the stimulus component produced a significant increase in lever-pressing resulting in presentation of the respective reinforcer-associated stimulus. Removal of cocaine and the associated stimulus during all schedule components led to extinction of drug-seeking behavior within six days. Lever-pressing resulting in presentation of the drug-associated stimulus was subsequently by non-contingent delivery of cocaine, but not by non-contingent presentation of the stimulus alone. These results suggest that different aspects of cocaine-seeking behavior, such as operant responding resulting in presentation of a cocaine-associated external stimulus, time to extinction in the absence of primary reinforcement, and reinstatement of responding following extinction can be measured in the rat. These tests may provide useful tools for the assessment of potential treatment drugs for human cocaine abuse.