During the last decades, large investments were made towards the development of numerical models and methods to forecast and analyze the different aspects of oil and gas recovery. In this context, modern simulators must be able to incorporate a wide variety of options to answer questions related to reservoir management accurately and effectively. In this work, we present a reservoir simulator based on a plugin architecture, where different formulations, solvers, and models can be developed, extended, and enhanced. With this approach, we use the black-oil model to implement traditional and state-of-the-art techniques, including fully- and adaptiveimplicit methods, heuristic and PID time-step controllers, Newton-Raphson and Inexact Newton, and C1-continuous and conventional phase-potential single-point upstream weighting. Several plugin configurations were tested and validated with commercial simulators, and their performances were used to determine which are the most suitable to solve multiphase flow problems.