Our report of a patient with severe tardive dyskinesia (TD) who has been exposed to both typical antipsychotic and clozapine, olanzapine and quetiapine during a 124-week follow-up period supports the possible beneficial effect of atypical antipsychotics on pre-existing symptoms of TD. Persistently high AIMS scores during all the periods of treatment with typical antipsychotics contrast strongly with the drop in scores that occurres in strict chronological sequence after switching to both clozapine (45%), olanzapine (27.8%) and quetiapine (85%). Since the reversal to haloperidol from the three atypical agents was systemically associated with a return to high AIMS scores, it seems likely that the improvement noted with clozapine, olanzapine and quetiapine represents a temporary symptomatic effect rather than a sustained resolution of the disorder. The olanzapine–clozapine–quetiapine rank order of increasing effectiveness against TD symptoms suggests that this property, although shared by the atypical antipsychotics, is to some degree drug-specific. Patient- and/or drug-dependent mechanisms may be involved in this gradient of effect.