The paper presents a simplified implementation of the Arakawa and Schubert (1974) cumulus parameterization, known as the Relaxed Arakawa–Schubert (RAS) scheme. The main simplifications in RAS include modifying the entrainment relation to avoid the costly calculation of the entrainment parameter and relaxing the state toward equilibrium rather than requiring it to be balanced each time the parameterization is invoked. The RAS scheme is evaluated using offline tests with the Global Atmospheric Research Program (GARP) Atlantic Tropical Experiment (GATE) Phase III data. Results show that RAS is equivalent to the standard Arakawa–Schubert implementation but is more economical and simpler to code. RAS also avoids the ill-posed problem that occurs in the standard implementation due to the need to solve for a balanced state. The paper discusses the continuous form of the parameterization, including the cloud model, cloud work function, and cumulus effects on large-scale budgets. Sensitivity tests are performed to explore the impact of the relaxation parameter and the order of cloud type selection. The results show that RAS produces solutions with nonzero mass fluxes for all possible cloud types, while the standard implementation overadjusts most cloud types. A semiprognostic evaluation using GATE Phase III data demonstrates that RAS successfully predicts precipitation rates, particularly during disturbed conditions. The computational efficiency of RAS is also compared with the standard implementation, showing that RAS is approximately four times faster.The paper presents a simplified implementation of the Arakawa and Schubert (1974) cumulus parameterization, known as the Relaxed Arakawa–Schubert (RAS) scheme. The main simplifications in RAS include modifying the entrainment relation to avoid the costly calculation of the entrainment parameter and relaxing the state toward equilibrium rather than requiring it to be balanced each time the parameterization is invoked. The RAS scheme is evaluated using offline tests with the Global Atmospheric Research Program (GARP) Atlantic Tropical Experiment (GATE) Phase III data. Results show that RAS is equivalent to the standard Arakawa–Schubert implementation but is more economical and simpler to code. RAS also avoids the ill-posed problem that occurs in the standard implementation due to the need to solve for a balanced state. The paper discusses the continuous form of the parameterization, including the cloud model, cloud work function, and cumulus effects on large-scale budgets. Sensitivity tests are performed to explore the impact of the relaxation parameter and the order of cloud type selection. The results show that RAS produces solutions with nonzero mass fluxes for all possible cloud types, while the standard implementation overadjusts most cloud types. A semiprognostic evaluation using GATE Phase III data demonstrates that RAS successfully predicts precipitation rates, particularly during disturbed conditions. The computational efficiency of RAS is also compared with the standard implementation, showing that RAS is approximately four times faster.