Machine learning classifications of first-episode psychosis (FEP) using neuroimaging have predominantly analyzed brain volumes. Some studies examined cortical thickness, but most of them have used parcellation approaches with data from single sites, which limits claims of generalizability. To address these limitations, we conducted a large-scale, multi-site analysis of cortical thickness comparing parcellations and vertex-wise approaches. By leveraging the multi-site nature of the study, we further investigated how different demographical and site-dependent variables affected predictions. Finally, we assessed relationships between predictions and clinical variables. 428 subjects (147 females, mean age 27.14) with FEP and 448 (230 females, mean age 27.06) healthy controls were enrolled in 8 centers by the ClassiFEP group. All subjects underwent a structural MRI and were clinically assessed. Cortical thickness parcellation (68 areas) and full cortical maps (20,484 vertices) were extracted. Linear Support Vector Machine was used for classification within a repeated nested cross-validation framework. Vertex-wise thickness maps outperformed parcellation-based methods with a balanced accuracy of 66.2% and an Area Under the Curve of 72%. By stratifying our sample for MRI scanner, we increased generalizability across sites. Temporal brain areas resulted as the most influential in the classification. The predictive decision scores significantly correlated with age at onset, duration of treatment, and positive symptoms. In conclusion, although far from the threshold of clinical relevance, temporal cortical thickness proved to classify between FEP subjects and healthy individuals. The assessment of site-dependent variables permitted an increase in the across-site generalizability, thus attempting to address an important machine learning limitation.