We propose a pervasive mobility classification method using radio beacons such as Global System for Mobile communications (GSM) and Wi-Fi traces. The model adopts different classifiers depending on the densities of radio beacons in different environments. We demonstrate how coarser-grained mobility states such as being stationary, walking, or driving can be satisfactorily inferred from our method using a data set of i) five hours gathered from one user in five differently-characterized areas and ii) sixteen hours gathered from sixteen individuals. Our model works across environments having different radio densities by employing the GSM-based classifier when Wi-Fi densities are too sparse with 81.54%. We also present that our model trained with the small data set gathered from one user is effectively applied to sixteen other individuals with 78% accuracy, which suggests the scalability of our model to new users.