The role of historical factors in establishing patterns of diversity in tropical mountains has been of interest for some time. In these regions, the historical processes of range fragmentation and contraction followed by dispersal are thought to be mediated by the interplay between rugged relief and climate fluctuations and likely explain most of the dynamics of diversification in plants and animals. Although empirical studies addressing the interaction between climate and topography have provided invaluable insights into population divergence and speciation patterns in tropical montane organisms, a more detailed and robust test of such processes in an explicit spatio-temporal framework is still lacking. Consequently, our ability to gain insights into historical range shifts over time and the genomic footprint left by them is limited. Here we used niche modeling and subgenomic population-level datasets to explore the evolution of two species of warbling finches disjunctly distributed across the Montane Atlantic Forest, a Neotropical region with complex geological and environmental histories. Population structure inferences suggest a scenario of three genetically differentiated populations, which are congruent with both geography and phenotypic variation. Demographic simulations support asynchronous isolation of these populations as recently as ~40 kya, relatively stable population sizes over recent time, and past gene flow subsequent to divergence. Throughout the last 800 ky, niche models predicted extensive expansion into lowland areas with increasing overlap of species distributions during glacial periods, with prominent retractions and isolation into higher altitudes during interglacials, which are in line with signs of introgression of currently isolated populations. These results support a dual role of cyclical climatic changes: population divergence and persistence in mountain tops during warm periods followed by periods of expansion and admixture in lower elevations during cold periods. Shallow genetic divergence could be associated with undiscovered cryptic diversity in this global biodiversity hotspot.