The nature and extent of face-processing impairments in individuals with autism spectrum disorder (ASD) remain contentious. The aim of this research study is to assess the face- and object-processing performance of individuals with ASD compared with typically developing controls. Our hypothesis was that individuals with ASD would be significantly impaired on tests of face processing but show intact object processing. More specifically, we tested two competing hypotheses to explain face-processing deficits: holistic hypothesis; second-order configural hypothesis. Twenty-six able adults with ASD and 26 intelligence quotient-matched typically developing controls completed two computerized tests of face and object discrimination. In task 1, the first picture (faces or cars) in a pair was presented as quickly as 40 msec to test holistic processing. In task 2, the decision was whether pairs of faces or houses had been altered in terms of the features or the distance between the features (the second-order configural properties). Individuals with ASD were impaired on all tests of face processing but showed intact object processing and the pattern of findings favored the holistic hypothesis. The heterogeneous pattern of performance in the clinical group showed that some individuals with ASD perform similarly to typically developing individuals in their face-processing skills, whereas others are more accurate in object processing compared with face processing.