On the heels of clozapine, we now have a number of newer agents (risperidone, olanzapine, quetiapine, sertindole, and ziprasidone). Are they all the same? What are the differences? How do we best understand them? In this article we review current clinical evidence to compare these issues on four measures of atypicality: EPS, prolactin elevation, superior efficacy in refractory/positive symptoms and efficacy against negative symptoms. All the newer agents are superior on EPS and, with the exception of risperidone, avoid prolactin elevation. Clozapine shows the most convincing efficacy in refractory schizophrenia, although comparative data concerning risperidone's benefit in this respect are also emerging. It is unclear, however, whether any of the agents produce a greater effect than conventional antipsychotics against positive symptoms in responsive patients. Both clozapine and olanzapine have demonstrated superior efficacy against negative symptoms, although it remains controversial whether this is an effect on primary or secondary symptoms. The precise pharmacologic mechanisms underlying "atypicality" remain unclear, but several conceptual frameworks are highlighted that characterize, and perhaps differentiate, these newer agents.