Apatite thermochronology is, in principle, uniquely suited to document the Cenozoic erosion of the Colorado Plateau (southwestern United States) and settle generations of debate regarding the region’s history of uplift, erosion, and fluvial incision. The protracted near-surface history of the Colorado Plateau bedrock, however, complicates the temperature sensitivity of apatite thermochronometers. This has confounded efforts to see clear evidence of late Cenozoic erosion, especially in the central Colorado Plateau, where this problem is compounded by the diverse detrital apatite grains in the region’s sedimentary bedrock. We overcome this problem in the thermal aureole of the Oligocene Henry Mountains intrusive complex, where these sandstones have apatite (U-Th-Sm)/He ages younger than 26 Ma with positive-slope age–effective U trends (3–25 Ma, 5–180 ppm eU) that resolve a distinctive late Cenozoic history. Thermal modeling results strongly suggest that the central Colorado Plateau was a stable Miocene landscape that was rapidly exhumed ∼1.5–2 km during the past 5 m.y., likely in the past 3–2 m.y., at time-averaged rates of ∼250–700 m/m.y. This demonstrates that substantial late Cenozoic erosion of the north-central plateau interior postdates the ca. 5.6 Ma integration of the Colorado River that lowered regional base level.