There are two interesting documents on the topic.
One is this review of central bank models written by Bank of Japan economists in Jan 2009. Hence, it misses all the recent criticisms on macroeconomic models doing rounds now. Second, is this superb article in WSJ which looks at how three professors (only one economics one) are trying to build better macro models for future.
First the BoJ review which looks at what kinds of models central banks use (suite of models) and focuses on BoJ models. It explains models through Pagan frontier (Pagan is an economist at Bank of England). Pagan Curve looks at two variables – Theoretical coherence and Empirical coherence. So you have a frontier on which all these models lie. Some are more inclined towards theory some towards empirical work. Ideal model should try to balance these two objectives. Very neat way of analysing. Based on this you have following models: