We propose a model for describing and predicting the performance of parallel numerical multilevel software on distributed memory architectures within a multi-cluster environment. The goal of the model is to allow reliable predictions to be made as to the execution time of a given code on a large number of processors of a given parallel system, and on a combination of systems, by only benchmarking the code on small numbers of processors. In addition to reporting on the performance of our model as applied to benchmark multigrid codes, we also provide results which demonstrate that this performance is also achieved when applied to a practical multilevel engineering code.