This paper presents a simple and accurate analytical model to dimension mobile cellular networks with QoS differentiation. QoS per applicative flow is commonly defined in GPRS/EDGE and 3G systems where streaming applications with real time properties and elastic data applications have to share radio resources. The need for accurate and fast-computing tools is of primary importance to tackle complex and exhaustive dimensioning issues. In this paper, we present a generic QoS modeling methodology developed in the context of EDGE networks but that can be adapted easily to a different technology. We develop a ldquoconditional product-formrdquo Markovian model that takes into account the QoS differentiation between real time and non-real time classes and gives closed form expressions for all the required performance parameters at a click speed. We compare our model with simulation and show its accuracy. The speed of our analytical tool allows tackling dimensioning issues that require several thousands of evaluations, which is not tractable with any simulation tool.