An early understanding of key metabolites of drugs is crucial in drug discovery and development. As a result, several in vitro models typically derived from liver are frequently used to study drug metabolism. It is presumed that these in vitro systems provide an accurate view of the potential in vivo metabolites and metabolic pathways. However, no formal analysis has been conducted to validate their use. The goal of the present study was to conduct a comprehensive analysis to assess if the three commonly used in vitro systems, pooled human liver microsomes, liver S-9 fraction, and hepatocytes, adequately predict in vivo metabolic profiles for drugs. The second objective was to compare the overall capabilities of these three systems to generate in vivo metabolic profiles. Twenty-seven compounds in the Pfizer database and 21 additional commercially available compounds of diverse structure and routes of metabolism for which the human ADME data was available were analyzed in this study to assess the performance of the in vitro systems. The results suggested that all three systems reliably predicted human excretory and circulating metabolite profiles. Furthermore, the success in predicting primary metabolites and metabolic pathways was high (>70%), but the predictability of secondary metabolites was less reliable in the three systems. Thus, the analysis provides sufficient confidence in using in vitro systems to reliably produce primary in vivo human metabolites and supports their application in early discovery to identify metabolic spots for optimization of metabolic liabilities anticipated in humans in vivo. However, the in vitro systems cannot solely mitigate the risk of disproportionate circulating metabolites in humans and may need to be supplemented with metabolic profiling of plasma samples from first-in-human studies or early human radiolabeled studies.