The initial enthusiasm about the second-generation or atypical antipsychotic drugs soon changed into criticism and debate, culminating in the controversial CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness), CUtLASS (Cost Utility of the Latest Antipsychotic Drugs in Schizophrenia Study) and EUFEST (European First-Episode Schizophrenia Trial) effectiveness trials. This review summarizes the results of three recent meta-analyses that compared second-generation antipsychotics (SGAs) with placebo, with conventional antipsychotics, and with SGAs head-to-head. We compare the meta-analyses with previous reviews and put them in the perspective of CATIE, CUtLASS and EUFEST. The data show that the SGAs are not a homogeneous group and that this confusing classification should be abandoned. We find that, overall, the data are consistent but experts interpret the same results differently. The debate seems to be driven more by values than by data; some place an emphasis on cost, others focus on extrapyramidal side-effects (EPS), weight gain, or efficacy. In our opinion, the SGAs are not the breakthrough that industry would like to maintain. They have different properties, so a clinician may individualize a treatment plan to a given patient's problems, a decision that should be shared with the patient. However, these drugs are important contributions to treatment, and most psychiatrists, let alone patients, would probably not want to do without them.