Nowadays the weather forecasting and the majority of the studies aimed at understanding the atmosphere (and climate) dynamics rely on the employment of numerical models. Starting from observations of the current state of the atmosphere, numerical weather prediction (NWP) models allow us to forecast future meteorological conditions. Numerical simulations are performed via supercomputers with high-level computational capacity. These machines are able to process millions of variables at the same time providing a realistic, although simplified, representation of the natural processes. Most of the time, this simplified form of reality is accurate enough to represent properly a given phenomenon and thus provides a useful analogue of the natural world. However, what makes the use of these models controversial is the risk of misinterpreting their outputs or misjudging the extent of their applications. This article gives an overview of what numerical models are and how they work. Attention will also be focused on their strengths and limitations.