In the last years, ground-based and space observations are bringing a wealth of constraints on the chemical composition of exoplanet atmospheres. Chemical models are particularly useful because they provide the theoretical framework to interpret these observations and make new predictions. I will describe how such models are built and discuss some recent efforts aimed to provide a more accurate description of the atmospheric composition, with emphasis on the mutual influence of temperature and disequilibrium composition, the quantification of abundance uncertainties, and the variation of the composition as a function of longitude.