This study aims to correct for long-term vertical land motions at tide gauges (TG) by estimating high-accurate GPS vertical velocities at co-located stations (GPS@TG), useful for long-term sea-level change studies and satellite altimeter drift monitoring. Global Positioning System (GPS) data reanalyses are mandatory when aiming at the highest consistency of the estimated products for the whole data period. The University of La Rochelle Consortium (ULR) has carried out several GPS data reanalysis campaigns with an increasing tracking network, an improving processing strategy and the best methodology. The geodetic results from the latest GPS velocity field estimated at ULR (named ULR5) are presented here. The velocity field includes 326 globally distributed GPS stations, from which 200 are GPS@TG (30% more than previous studies). The new GPS data processing strategy, the terrestrial frame definition and the velocity estimation procedures are described. The quality of the estimated vertical velocities is empirically assessed through internal and external velocity comparisons, including the analysis of the time-correlated noise content of the position time series, to be better than 0.6 mm/yr (2 sigma). The application of this velocity field is illustrated to appraise to what extent vertical land motions contaminate the estimates of satellite altimetry drifts. The impact on the altimeter-derived sea level trends was evaluated to be up to 0.6 mm/yr. Worldwide TGs were grouped into regions in order to explore long-term spatial sea level variability in the rates of sea level change. By taking into account the vertical land motion of the tide gauges, the dispersion of the observed sea level rates within each region was reduced by 60%. Long-term regional mean sea level variations up to 70% from the global mean were found.