Background This study introduces a laboratory model of compulsive behavior based on persistence in the context of rewarded spatial alternation. Methods Rats were screened for spontaneous persistence during T-maze reinforced alternation. Experiment 1: One high and one low spontaneous persistence group ( n = 8) received 20 injections of fluoxetine, a matched pair saline, both followed by 4 days of meta-chlorophenylpiperazine (mCPP) challenge. Experiment 2: Five matched groups of rats ( n = 9) received pretreatment (20 injections) with fluoxetine, mCPP, desipramine, diazepam or saline, followed by 4 days of mCPP challenge (fluoxetine in mCPP group). After washout, animals received 2 days of naratriptan, followed by another 2-day mCPP challenge. Results In both experiments mCPP significantly increased persistence in saline controls. Fluoxetine also acutely increased persistence scores: after a gradual return to baseline, these scores showed tolerance to mCPP. Experiment 1: This pattern was significant in high but not low initial persistence groups. Experiment 2: Fluoxetine and mCPP showed cross-tolerance. Neither desipramine nor diazepam protected against mCPP challenge. Persistence scores returned to baseline during washout and naratriptan and were thereafter increased by another mCPP challenge in all but the fluoxetine and mCPP groups, suggesting 5-HT 2C receptor mediation. Conclusions This model is based on spontaneous persistence behavior showing pharmacological responses concordant with those of compulsive symptomatology.