For about two decades, SWEPOS (the Swedish Permanent GNSS network) pillar stations have been used indifferent geodetic and geodynamic studies. To keep continuous measurements of these long lived pillar stationsand at the same time modernizing the SWEPOS network, it has been decided to install new truss mast stations,equipped with modern and individually calibrated antennas and radomes, capable of tracking all new GNSSsatellites. Installation of mast stations started in 2011. Today, each pillar station in the SWEPOS permanent GNSSnetwork has a close-by truss mast station, mostly in 10 meters distance with individual calibrated Leica chokering antenna and its attachment (LEIAR25.R3, LEIT). Due to their closeness to pillars, the modern mast stationsmay provide additional information for the analysis of ground movements in Sweden e.g. to distinguish betweentectonic and geodynamic processes (e.g. land uplift in Sweden).In this study, we have used two datasets from two different seasons for 21 pillars and 21 mast stations andformed different networks. The mast network has been processed using both IGS standard (type) and individuallycalibrated PCV (Phase Center Variation) models and therefore the effect of these two different PCV models onheight components has been investigated. In a combined network, we processed all 42 stations (21 pillars+21mast) to see how this multi-baseline network (861 baselines) combination differs from independent mast or pillarnetworks with much less baselines (210 baselines). For our analysis, we used the GAMIT-GLOBK softwareand compared different networks. Ambiguity resolutions, daily coordinate repeatability and differences betweenheight components in different solutions are presented. Moreover, the GAMIT and BERNESE solutions forcombined mast and pillar networks are compared.Our results suggest that the SWEPOS truss mast stations can reliably be used for crustal deformation studies.The comparison between pillar and mast stations shows similar time series for different horizontal and verticalcomponents and their Normalized rms (nrms) and weighted rms (wmrs) are almost equal.Comparison of standard and calibrated PCV models for mast stations show notable differences in height compo-nents and reach up to14 mm. These differences are antenna-dependent and are not systematic offsets. Therefore,whenever available, individual calibrated antenna models have to be used instead of standard (type) calibratedmodels.This study is part of the Swedish CLOSE III research project between Lantmäteriet, SP, and Chalmers Universityof Technology.