The challenge of continuously improving engine fuel economy and emissions has pushed the automotive industry to adopt more efficient procedures for modeling, system simulation and model-based control. Such procedures require accurate and computationally efficient engine system simulation models that are able to predict the cylinder charge composition and the thermodynamic conditions. To this extent, reduced-order models for unsteady compressible flow systems, namely models derived from the fundamental conservation laws without over-simplifying the topology of the engine air path, are receiving considerable interest for performance simulation and control design.This paper presents a comparative study on modeling approaches for the prediction of one-dimensional unsteady compressible flow in the internal combustion engine air path system. Specifically, the paper compares a well-known second-order, shock-capturing finite-difference scheme with two novel solution methods, namely a finite-volume method and a model-order reduction method. The study aims at evaluating the ability of each method to trade-off the prediction accuracy and robustness with the computation time, in light of potential applications to real-time simulation. This is achieved by increasing the discretization length of each method, and evaluating the accuracy of the predicted response in the time and frequency domain.The characterization of the intake and exhaust flows in the manifolds of a single-cylinder engine is considered as a case study. The results are compared for the three solution methods at different discretization lengths to evaluate the ability of each model to retain accuracy and robustness. The computation times of the different methods are also evaluated.The results presented in this paper establish a clear trade-off between accuracy, stability and computation time for each solution method. This allows one to formulate considerations on the advantages and disadvantages of each method in light of potential applications to engine performance prediction, optimization and control design.