In experiment 1, two different strains of mice [C57BL/6J (B6) and DBA/2J (D2)] were allowed to nosepoke for 5 microliters intravenous (IV) infusions during 2-h daily sessions. Two nosepoke holes were available, only one of which was reinforced on an FR-3 schedule with a 10-s time-out indicated by a light inside the reinforced nosepoke hole. During the first nine sessions, infusions were saline. On subsequent sessions, mice acquired nosepoking for 0.5 mg/kg cocaine. Finally, all mice were extinguished by again receiving only saline infusions. Cocaine acted as a reinforcer in both strains. In experiment 2, different mice from the same two strains were allowed to acquire nosepoking for IV cocaine at one of three unit doses (0.5, 1.0, or 2.0 mg/kg). Although there were no effects of unit dose on rate of acquisition, B6 mice were faster in acquiring self-administration behavior than were D2 mice. Experiment 3 assessed behavior in the same mice, after acquisition had occurred. D2 mice nosepoked at a lower rate at asymptote than did B6 mice, but with a higher preference for the cocaine reinforced hole. Unit doses of cocaine were then manipulated within subjects, from 0.125 to 2.0 mg/kg per infusion. Higher doses yielded lower response rates than lower doses, both between and within subjects. Behavior in D2 mice relative to B6 mice also appeared to be shifted to the left of the dose-response curve measured within-subjects. Together, these findings indicate that although cocaine serves as a reinforcer in both strains, there are genetic differences in the pattern of cocaine self-administration between these two mouse strains.