The article evaluates the accuracy of six small-molecule force fields (OpenFF Parsley, Sage, GAFF, CGenFF, and OPLS3e) in predicting protein-ligand binding affinities using alchemical relative binding free energy (RBFE) calculations on 598 ligands and 22 protein targets. While OpenFF Parsley, Sage, GAFF, and CGenFF show comparable accuracy, OPLS3e is significantly more accurate. A consensus approach using Sage, GAFF, and CGenFF achieves accuracy comparable to OPLS3e. The accuracy of predictions depends on force field parameters, input preparation, and sampling convergence. Large perturbations and non-converged simulations lead to less accurate predictions. The input structures, Gromacs force field files, and analysis Python notebooks are available on GitHub. The study highlights the importance of accurate force field parameters in predicting binding affinities. The OpenFF toolkit was used to parameterize ligands with AM1-BCC charges. GAFF parameters were assigned using Antechamber and ACPYPE. CGenFF/MATCH* parameterization involved atom typing with MATCH and replacing bonded parameters with CGenFF v3.0. OPLS3e was used in FEP+ results, and the consensus approach combined results from multiple force fields. The results show that OPLS3e has the lowest RMSE (1.3 kcal/mol) and the consensus approach performs similarly. The accuracy of predictions is influenced by factors such as the convergence of simulations, the number of heavy atoms, and the complexity of the perturbations. The study also demonstrates that improvements in force field parameters can significantly enhance the accuracy of binding affinity predictions. The OpenFF-2.0 version showed improved accuracy compared to OpenFF-1.0, particularly for ester groups and hydroxyl groups. The study concludes that while public force fields perform comparably, OPLS3e and the consensus approach offer higher accuracy. The results emphasize the importance of proper input preparation, sampling convergence, and the use of accurate force field parameters in predicting binding affinities.The article evaluates the accuracy of six small-molecule force fields (OpenFF Parsley, Sage, GAFF, CGenFF, and OPLS3e) in predicting protein-ligand binding affinities using alchemical relative binding free energy (RBFE) calculations on 598 ligands and 22 protein targets. While OpenFF Parsley, Sage, GAFF, and CGenFF show comparable accuracy, OPLS3e is significantly more accurate. A consensus approach using Sage, GAFF, and CGenFF achieves accuracy comparable to OPLS3e. The accuracy of predictions depends on force field parameters, input preparation, and sampling convergence. Large perturbations and non-converged simulations lead to less accurate predictions. The input structures, Gromacs force field files, and analysis Python notebooks are available on GitHub. The study highlights the importance of accurate force field parameters in predicting binding affinities. The OpenFF toolkit was used to parameterize ligands with AM1-BCC charges. GAFF parameters were assigned using Antechamber and ACPYPE. CGenFF/MATCH* parameterization involved atom typing with MATCH and replacing bonded parameters with CGenFF v3.0. OPLS3e was used in FEP+ results, and the consensus approach combined results from multiple force fields. The results show that OPLS3e has the lowest RMSE (1.3 kcal/mol) and the consensus approach performs similarly. The accuracy of predictions is influenced by factors such as the convergence of simulations, the number of heavy atoms, and the complexity of the perturbations. The study also demonstrates that improvements in force field parameters can significantly enhance the accuracy of binding affinity predictions. The OpenFF-2.0 version showed improved accuracy compared to OpenFF-1.0, particularly for ester groups and hydroxyl groups. The study concludes that while public force fields perform comparably, OPLS3e and the consensus approach offer higher accuracy. The results emphasize the importance of proper input preparation, sampling convergence, and the use of accurate force field parameters in predicting binding affinities.