We investigated the efficacy and safety of aripiprazole in first-episode psychosis and explored the association between early response and later response to this medication. This was a 6-week, open-label, multicenter trial. The study population consisted of 59 patients with a DSM-IV diagnosis of a schizophreniform disorder, schizoaffective disorder, schizophrenia, or psychotic disorder not otherwise specified. The primary outcome measures were the Positive and Negative Syndrome Scale (PANSS) and the Clinical Global Impression-Severity scale. To assess the safety, we measured the drug-related adverse events, weight, and lipid-related variables. The sensitivity, specificity, positive predictive value, and negative predictive value were calculated for the response status at weeks 2 and 3 to predict the subsequent response at week 6. Among the 59 participants, 38 were able to complete the 6-week trial. Treatment with aripiprazole resulted in significant improvement in the PANSS and Clinical Global Impression scores over time. The response rate (defined as a ≥30% decrease in the PANSS total score from baseline to the last observation) was 69.1%. The most accurate prediction of later response in terms of negative predictive value and specificity was a reduction in the PANSS total score from baseline to week 3 of at least 20%. Aripiprazole had a modest side effect burden and was characterized by a safe profile with respect to weight and metabolic side effects. These results indicate that aripiprazole is effective and safe in the treatment of first-episode psychosis. The response at week 3, rather than week 2, predicted the later response more accurately.