The air spring is the main part of the secondary suspension of passenger railway vehicles. The aim of this paper is to review existing modelling techniques for air springs in order to check if challenges set in the past decade for available models have been met. The advantages and disadvantages of different air spring models (phenomenological/mechanical, thermodynamic, analytic, FEM) are summarised and discussed from the point of view of: model accuracy, multiphysics interaction, influence of structural and material non-linearities, obtention of parameters, frequency range and the balance between accuracy and computational effort. The first conclusion is that current research is mainly focused on the vertical behaviour with less attention paid to the lateral performance. Moreover, it is concluded that further research is needed to include non-linearities of the bellow and to consider fluid-structural interaction; this would allow improving the model of vertical behaviour and evaluating better the lateral performance of the pneumatic system. FEM models might be an interesting tool that allows performing a more complete analysis of air springs (combining different physics, including material non-linearities, considering the real shape of the bellow and reinforcing fibres, etc) favouring the comfort analysis and including the lateral dynamics of the air spring.