Spider monkeys (Ateles) are one of the most endangered groups of primates in the Neotropics. The genus is widely distributed from Mexico to the north of Bolivia and includes many morphologically distinct forms in terms of pelage color and patterning. The taxonomy, phylogenetic relationships, and biogeographic history of the genus have been subject to much debate, making scientific communication difficult and creating challenges for conservation actions. We extracted DNA from samples of all currently recognized species of spider monkeys collected from across the geographic range of the genus, sequenced ~3.5. kilobases of coding sequence from the mitochondrial genome, and used this large dataset to (a) infer the phylogenetic relationships among the different forms of spider monkeys, (b) evaluate whether currently recognized species of spider monkeys form reciprocally monophyletic groups that are concordant with contemporary classifications, and (c) estimate divergence dates among the different lineages of Ateles. We found that all proposed species of spider monkeys for which we have samples from multiple localities indeed appear to form monophyletic groups. However, in contrast to previous studies, several of our analyses robustly inferred Ateles marginatus from northeast Brazil as the sister taxon to all other spider monkeys. A Bayesian dating analysis suggests that the most recent common ancestor of extant Ateles dates to ~6.7. Ma, in the late Miocene, and most species-level splits within the genus took place in the late Pliocene, suggesting that the modern diversity in spider monkeys cannot be explained principally by isolation and divergence of populations in forest refugia during the Pleistocene. Based on our new phylogenetic inference and dating analysis, we propose a revised biogeographic scenario for the evolution of this genus.