Most currency option pricing models, such as the Garmen-Kohlhagen model and Grabbe model, assume that changes of exchange rates are normally distributed. The assumption is not consistent with empirical evidence and may create biased volatility estimates and unreliable hedged ratios. The author recommends the McCullouch model which can capture the extreme changes in foreign exchange rates and thus may be useful for pricing currency options during crisis such as the stock market crash in 1989, the crisis of the currencies of the European Monetary system at the end of 1992 as well as the 2008 sub-prime mortgage loan and banking crisis. The author will also evaluate the performance of the models.