In addition to what others have said, it might help you to know that Monte Carlo methods, at least in the case of simulating rarefied gas flows (Direct Simulation Monte Carlo - DSMC), perform much, much better the further you are from equilibrium. Situations with low speeds / low pressure differences / low temperature differences usually require a prohibitively high computational time, unless you are using a heavily modified version of DSMC. The reason is that the statistical noise hides the (very low) gradients of pressure, temperature, etc, so you need to have a sample of many particles or average over a long enough time interval.
I do not have experience with MD, so please take my opinion with a grain of salt, but I do not see a reason for the method not being able to simulate both types of systems in theory, assuming that you confine your investigation in a very small area and time window. But I suppose you will have better performance in some cases than others.