The analytical procedures for all-way stop-controlled intersections in the 2000 edition of the "Highway Capacity Manual" (HCM) lack a model to estimate the 95th percentile queue length. This is considered a major shortcoming of the procedures, since queue length is one of the key parameters used in the traffic engineering design process. This paper provides an assessment of three queue length models for potential application at all-way stop-controlled intersections. One is the theoretically based queue length model included in the HCM for two-way stop-controlled intersections, and the others are empirically based models: a model based on simulation and a newly calibrated model based on field data. The three models were compared on the basis of field data from 18 sites. All the models produce queue length results that closely match the field data. The queue length model for two-way stop-controlled intersections was developed with a sound theoretical basis; therefore, it is recommended that this model be adopted for analyzing all-way stop-controlled intersections in the HCM. However, the empirical models are easy to use and proved to be of the same level of accuracy. The empirical models can be used as an alternative, especially when a quick estimation of the queue length is desired.