Abstract: Objectives: On 7 January 2025, at 09:05:16 Beijing time, an Ms 6.8 earthquake occurred in Dingri, Xizang, China. It was located in a tectonically complex region near the intersection of the Xainza-Dinggye Fault and the South Xizang Detachment System, and its principal rupture zone extended to the surface. This event caused serious damage and provided an important case for investigating the fault slip distribution, rupture mechanism, and regional deformation processes. Therefore, this study aims to obtain a high-precision three-dimensional coseismic deformation field of the 7 January 2025 Dingri (Xizang) Ms 6.8 earthquake from high-rate GNSS observations, and to characterize rupture features by inverting fault geometry and slip distribution based on the derived deformation field. Methods: High-rate (1 Hz) GNSS observations from seven stations within 300 km of the epicentral area were processed using the Precise Point Positioning with Ambiguity Resolution (PPP-AR) strategy to obtain high-frequency coordinate time series, which were then transformed into local east, north, and up (ENU) displacement time series for coseismic deformation analysis. Considering the relatively low signal-to-noise ratios of the up-component displacement time series at five stations, namely XZNL, XZJL, XZJZ, XZSG, and XZSZ, these vertical displacement time series were further denoised before coseismic displacement estimation. Specifically, the selected vertical displacement time series were first decomposed into a series of intrinsic mode functions using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN). A Multiscale Permutation Entropy (MPE)-based criterion was then applied to identify noise-dominated components, and the effective displacement signals were reconstructed by retaining the components related to coseismic deformation. The three-dimensional coseismic deformation field was subsequently estimated by differencing the pre-event and post-event mean positions. Based on previous studies, the fault strike and dip were constrained, while the fault length, other geometric parameters, and slip distribution were estimated using the Okada elastic dislocation model. A two-step inversion strategy was adopted. In the first step, the fault geometry parameters were inverted using a uniform-slip dislocation model and the Particle Swarm Optimization (PSO) algorithm. In the second step, the slip distribution was estimated using a distributed-slip model and the Steepest Descent Method (SDM). The consistency and reliability of the inversion results were further evaluated by comparing the observed and modeled three-dimensional displacements. Results: The CEEMDAN-MPE denoising scheme effectively improved the quality of the selected vertical displacement time series. The signal-to-noise ratio was increased by up to approximately a factor of two, with the largest improvement observed at station XZJL. The derived coseismic deformation field shows that the deformation was dominated by the horizontal component, was primarily concentrated in the near field, and exhibited pronounced east-west asymmetry. The maximum horizontal displacement reached 89.25 mm, with a standard error of 1.88 mm, while the maximum vertical displacement reached 14.85 mm, with a standard error of 7.06 mm. The seismogenic fault has a strike of 189.33° , a dip of 52.8° , a rake of −61.06° , and a fault length of 50.68 km, indicating predominantly normal faulting with a strike-slip component. Slip was mainly concentrated at depth of 0-12 km, with a peak slip of 5.86 m. The estimated moment magnitude was Mw 7.15, which is generally consistent with previous studies. Conclusions: High-rate GNSS PPP-AR strategy combined with CEEMDAN-MPE denoising enables robust retrieval of three-dimensional coseismic deformation and supports reliable inversion of fault geometry and slip distribution constrained by the Okada elastic dislocation model. Although only two near-field GNSS stations were available in this study, they played a key role in constraining shallow deformation. The results indicate that even limited near-field high-rate GNSS observations can effectively capture shallow slip information. Therefore, this study provides a methodological reference for rapidly obtaining the macroscopic rupture characteristics of the seismic source in rapid-response scenarios. The existing GNSS stations within 300 km of Dingri County demonstrate the potential of regional high-rate GNSS networks for coseismic deformation monitoring and fault slip distribution inversion of strong regional earthquakes with magnitudes of 6.0 and above.