Recent experiments suggest that GABA compounds produce a clinically relevant modulation of cocaine reinforcement. This review summarizes the results of a number of studies that examined the effect of the GABAB agonist baclofen on cocaine self-administration using a variety of schedules of reinforcement. The results demonstrate that baclofen causes a dramatic reduction in cocaine self-administration which does not appear to be accounted for by a general disruption of behavior. However, the effect is dependent on the unit injection dose of cocaine and on the response requirements of the schedule. These results predict that in a clinical setting any potential therapeutic effect of baclofen will interact with the cost and availability of cocaine.