Aim Ancient tropical mountains are megadiverse, yet little is known about the distribution of their species. We aimed to disentangle the effects of latitudinal and elevational gradients on the distribution of species of Aculeata and to understand the effects of climatic variables across different spatial scales of diversity (alpha-, gamma-, and beta-diversity). Location Campo rupestre in the Espinhaco Mountain Range, Southeast Brazil. Taxon Bees, wasps, and ants (Aculeata: Hymenoptera). Methods We used a unique dataset built from sampling species of Aculeata at 24 study sites across 12 mountains, covering 1200 km from south to north and an elevational range of 1000 to 2000 m. We explored the elevational and latitudinal patterns of alpha- (site), gamma- (mountain), and beta-diversity among samples at each location (beta(Local)). We also tested the effect of elevational range on beta-diversity in each mountain (beta(Mountain)) and, on a larger scale (beta(Regional)), if beta-diversity is influenced by geographical and environmental distances. Finally, we tested whether climatic variables underpin the observed patterns. Results Latitude had no effect on diversity. We found a decrease in both site and mountain diversity and, only for bees, beta(Local) increased with elevation. Climatic variables (temperature, wind, and precipitation) and their interactions were important drivers of diversity, with temperature being the most important. Finally, beta(Mountain) increased with mountain elevation range, and beta(Regional) increased with the geographical and environmental distances. Main conclusions Our results showed that variation in species richness and composition across mountains is strongly associated with elevational gradient, which showed stronger climatic variation than latitudinal gradient. Therefore, despite having narrow elevational ranges, the biogeographical effects of tropical mountains drive high diversity. Facing global climate changes, this limited elevational gradient may limit species range shifts, leading to severe biodiversity losses.