Miocene vertebrate localities are uncommon in central South America. In Bolivia, the best known mammalian faunas of this period come from Quebrada Honda (late middle Miocene, Tarija Department) and Cerdas (early middle Miocene, Potosí Department). The Achiri locality (La Paz Department) was reported first by Hoffstetter in 1972. Subsequently, campaigns were conducted in this locality by several paleontologists, including Villarroel, Anaya, Saint-André, and by our team over the last decade. Recently, we have obtained two precise absolute dates (40Ar/39Ar) on feldspar contained in ashes intercalated between fossiliferous levels and confirmed a late Miocene age (10.35±0.07 Ma and 10.42±0.09 Ma, late Mayoan–early Chasicoan South American Land Mammal Ages) as suggested by Marshall and colleagues in 1983. Almost all the specimens come from the Cerros Virgen Pata and Jiska/Jacha Pisakeri localities, the latter located 3–4 km southeast of Achiri village. In the past, discoveries of numerous spectacular specimens have allowed the identification of new mammalian species such as the notoungulates Plesiotypotherium achirense and Hoffstetterius imperator, the sparassodontan Borhyaenidium altiplanicum, and the xenarthrans Trachycalyptoides achirense and Xyophorus villarroeli. Xyophorus was erected by Ameghino in 1887 on the basis of a dentary fragment from the lower Miocene Santa Cruz Formation (Argentina). This taxon is generally considered to be a nothrotheriid sloth (although has never been formally included in a phylogenetic analysis based on osteological characters). Six species are recognized in Argentina. This genus is also recorded in Achiri through the endemic species X. villarroeli, and also in Cerdas and Quebrada Honda through X. cf. bondesioi. Unfortunately, all the specimens referred to Xyophorus are extremely fragmentary. Here we present a partial skull (MNHN-Bol-V 12690, National Museum of Natural History, La Paz, Bolivia) discovered in Achiri, belonging to an adult, and referred as “Xyophorus” villarroeli. It consists of a right posterolateral portion of the skull, including parts of squamosal, parietal, basioccipital, exoccipital, and its complete ear region with ectotympanic, entotympanic, and petrosal. Preliminary observations of this new specimen reveal the presence of at least seven autapomorphies, including a very rugose external surface of ectotympanic, a clear contact between styliform process of ectotympanic and pterygoid, and a reduced or absent subarcuate fossa. This megatherioid sloth shares several synapomorphies with nothrotheriids, including a dorsoventrally elongated ectotympanic and an ovate stylohyal fossa. It exhibits also transitional features between basal megatherioids and nothrotheriids, like a ventral portion of the ectotympanic that is expanded transversely in ventral view (more than Hapalops and less than Nothrotheriidae) and deeper in lateral view than that of Hapalops, although similar in proportions to Pronothrotherium and Mionothropus. This specimen thus suggests that “Xyophorus” villarroeli could be an early-diverging nothrothere, with affinities to Hapalops and also early Nothrotheriidae, and probably distinct from Xyophorus of more austral localities. A comprehensive phylogenetic analysis of the Megatherioidea including this form from Achiri, Aymaratherium from the early Pliocene of Pomata-Ayte, and Lakukullus and Hiskatherium from Quebrada Honda, should allow for a better understanding of the relationships among Patagonian and Andean Megatherioidea.