In this paper, we develop a new class that generates flexible optimal models. The appropriate features of the new class and a special member are studied by utilizing analytical, graphical, and numerical methodologies. For the estimation of unknown parameters, the maximum likelihood, least-square, and percentile methods are discussed and selection is based on bias and mean square error via an extensive simulation study. Five-life time data sets are evaluated, revealing that the new class has a significant advantage over well-known competitors.