Accurate simulation of molecules reacting on metal surfaces, which can help in improving heterogeneous catalysts, remains out of reach for several reactions. For example, a large disagreement between theory and experiment for HCl reacting on Au(111) still remains, despite many efforts. In this work, the dissociative chemisorption of HCl on Au(111) is investigated with a recently developed MGGA density functional (MS-RPBEl) and a high-dimensional neural network potential. Additionally, previous experimental sticking probabilities are re-examined. A considerably improved agreement between experiment and theory is obtained, although theory still overestimates experimental sticking probabilities by a factor of 2–7 at the highest incidence energy. Computed and measured vibrational transition probabilities are also in improved agreement. Several dynamical effects such as angular steering and energy transfer from the molecule to the surface are found to play an important role.