This study was designed to explore the relationship between craving and cocaine-seeking behavior with the use of both subjective and behavioral measures. Five males and five females who have used crack at least two times a week for 6 months, and who reported using 0.5 g of crack within 24 h on at least one occasion, participated in an inpatient study. Subjects underwent a total of four experimental sessions, during which they were exposed to either neutral (Neutral Stimuli Condition) or cocaine-related (Cocaine Stimuli Condition) external and internal stimuli. Subjects were exposed to each stimuli condition twice, on separate days, in randomized order. External stimuli comprised neutral or cocaine-related videotapes and paraphernalia, and the internal stimulus was either a 5-mg (`placebo') or 0.4 mg/kg delivery of cocaine. At baseline and after each stimulus exposure, subjects completed a composite cocaine craving questionnaire. Subjects next worked on concurrently-available fixed-ratio tasks either for tokens that could be exchanged for money ($2) or for tokens that were exchangeable for deliveries of cocaine (0.4 mg/kg). The results show that subjects reported significantly greater cocaine craving after exposure to cocaine-related vs. neutral stimuli, indicating that craving for cocaine can be successfully modeled in a laboratory setting. However, this change in subjective response did not predict drug-seeking behavior. The number of cocaine tokens earned following exposure to the cocaine-related vs neutral stimuli was similar. These results suggest that in a laboratory setting, craving may be unrelated to cocaine-seeking behavior in non-treatment-seeking cocaine users.