High-resolution, garnet-based pressure-temperature (P-T) paths were obtained for nine rocks across the Himalayan Main Central Thrust (MCT) (Marsyangdi River transect, central Nepal). Paths were created using garnet and whole rock compositions as input parameters into a semiautomated Gibbs free-energy-minimization technique. The conditions recorded by the paths, in general, yield similar T but lower P compared to estimates from mineral equilibria and quartz-in-garnet Raman barometry. The paths are used to modify a model based on a two-dimensional finite difference solution to the diffusion-advection equation. In this model, P-T paths recorded by the footwall garnets result from fault motion at specified times, thermal advection, and alteration of topography. The best fit between the high-resolution P-T paths and model predictions is that from 25 to 18 Ma, samples within the MCT footwall moved at 5 km/Ma, while those in the hanging wall moved at 10 km/Ma. Under these conditions, topography grew to 3.5 km. A pause in activity along the MCT between 18 and 15 Ma allows heat to advect and may be due to a transfer of tectonic activity to the structures closer to the Indian subcontinent. During this time, the topography erodes at a rate of 1.5 km/Ma. Thrusting within the MCT footwall reactivates between 8 and 2 Ma with exhumation rates up to 12 mm/yr since the Pliocene. The results suggest the potential for the highestresolution garnet-based P-T paths to record both the thermobarometric consequences of fault motion and large-scale erosion.Plain Language Summary The Main Central Thrust (MCT) is a major Himalayan fault system largely responsible for the generation of its high topography. Garnets across the MCT record their growth history in the crust through changes in their chemistry. These chemical changes can be extracted and modeled. Here we report detailed pressure-temperature paths recorded by garnets collected across the MCT along the Marsyangdi River in central Nepal. The paths track evolving conditions in the Earth's crust when the MCT was active during the growth of the Himalayas. The results suggest that the MCT formed as individual rock packages moved at distinct times. Further modeling makes predictions about how the Himalayas developed, including that the MCT may have ceased motion 18-15 million years ago, as other faults closer to the Indian subcontinent became active, and that it reactivated 8-2 million years ago, leading to the generation of high topography. The modeling also suggests that very high erosion rates occurred within the range after reactivation. Although garnets have long been used to understand how fault systems evolve, we provide details of an approach that allows higher-resolution data to be extracted from them and show how they could be used to track large-scale erosion.