Several methods for modelling and finite element analysis of tubular welded joints are described in various design codes. These codes provide specific recommendations for modelling of the welded joints, using simple weld geometries. In this paper, experimental hot-spot strain range results from a full-scale automatically welded K-node test are compared to corresponding finite element models. As part of investigating the automatically welded K-joint, 3D scans of the weld surfaces have been made. These scans are included in the FE models to determine the accuracy of the FE models. The results are compared to an FE model with a simple weld geometry based on common offshore design codes and a model without any modelled weld. The results show that the FE model with 3D scanned welds is more accurate than the two simple FE models. As the weld toe location of the 3D scanned weld is difficult to locate precisely in the FE model and as misplacement of strain gauges are possible, stochastic finite element modelling is performed to analyse the resulting probabilistic hot-spot stresses. The results show large standard deviations, showing the necessity to evaluate the hot-spot stress method when using 3D scanned welds.