Abstract
A new useful version of the Frechet model is introduced and studied. Some of its properties are derived. The method of maximum likelihood is used for estimating the unknown parameter via two real data applications. The new version is much better than other important competitive Frechet models in modeling two real data sets.