This study investigated early childhood corpus callosum development, a critical process for cognitive maturation and implicated in Autism Spectrum Disorder (ASD), using sex-specific growth curve models. Structural MRI data from 295 typically developing children (TDC; aged 1–6 years) were used to model age- and sex-dependent changes in ten morphometric parameters, including subregion volumes and midsagittal plane features. Analyses revealed nonlinear developmental trajectories, region-specific growth rates, and earlier developmental peaks in females. We applied these normative models to an independent dataset of 41 TDC and 26 children with ASD, acquired on a different scanner. Classifiers trained on deviations from the growth curves accurately distinguished children with ASD from TDC (mean Area Under the Receiver Operating Characteristic Curve [AUC] = 0.95), demonstrating model generalizability. These findings establish sex-specific corpus callosum growth curve models as a quantitative, generalizable tool for characterizing typical development and detecting atypical morphometry, offering a promising approach for early, objective ASD diagnosis and potentially facilitating timely intervention. Further study of model generalizability across more diverse populations is warranted.