Introduction of the National Institute of Mental Health's Research Domain Criteria and revision of diagnostic classification for Autism Spectrum Disorder in the latest diagnostic manual call for a new way of conceptualizing heterogeneous ASD features. We propose a novel conceptualization of ASD, borrowing from the schizophrenia literature in clustering ASD features along positive, negative, and cognitive dimensions. We argue that this dimensional conceptualization can offer improved ability to classify, diagnose, and treat, to apply and predict response to treatment, and to explore underlying neural and genetic alterations that may contribute to particular feature clusters. We suggest the proposed conceptualization can advance the field in a manner that may prove clinically and biologically useful for understanding and addressing heterogeneity within ASD.