The use of detailed chemical kinetics combined with complex three-dimensional computer aided design (3-D CAD) geometries makes it practically unaffordable to simulate combustion in real configurations such as domestic gas cooking burners. Global or reduced reaction mechanisms are alternatives to alleviate this high computational load, but they are generally formulated on certain operating ranges or assumptions. In this work, the performance of chemical reaction mechanisms to simulate methane combustion is analyzed comparing with available experimental measurements and detailed chemistry. Examined geometrical configurations are one-dimensional (1-D) premixed flame, two-dimensional (2-D) partially premixed flame, and 3-D domestic gas cooking burner. Laminar flame speed and variables of interest such as temperature, major species, and heat exchange are the targets for performance comparison. Results show that global mechanisms could be used in certain cases for getting overall temperature fields and thermal balances, but never to accurately predict laminar flame speed and pollutant emissions. Skeletal models predominantly present good behavior in all the analyzed cases, although some of them show slight deviations. Mechanisms are ranked based on their statistical evaluation of agreement with reference data. In general, results state that the more species and reactions included, the better accuracy obtained but the higher computational time required. Most skeletal mechanisms are good alternatives to detailed chemistry; therefore, the final choice will be mainly influenced by the availability of computational resources and the required target of the computational fluid dynamics (CFD) analysis.