Inferring mobility states such as being stationary, walking, or driving is critical for transportation studies, urban planning, health monitoring and epidemiology. Our goal is to build a pervasive mobility classification system using smartphones while focusing on large deployment, which poses new design requirements: low processing complexity, high energy efficiency and high user-time coverage. Previous work focused on fine-grained location-based mobility inference using global positioning system (GPS) data. However, GPS-based mobility characterization raises many issues, such as spotty coverage and battery drainage, that makes it inadequate to meet our application goals. In this paper, we propose a new mobility classification method using radio beacons such as Global System for Mobile communications (GSM) and Wi-Fi traces. This method enables mobility-based applications to provide users with ubiquitous services while using energy-inexpensive existing infrastructures. We demonstrate how coarser-grained mobility states can be satisfactorily inferred from our method using a data set of five hours gathered from one user in five differently- characterized areas with 80.2% precision and 80.3% recall.