A moving box model was used to compare formaldehyde (HCHO) predictions obtained with three photochemical oxidant mechanisms. In simulations using a typical urban mix of organic pollutants, the mechanisms were found to yield HCHO predictions that agreed to within 10% of each other provided the same reaction rates were used in all three mechanisms to describe the photolysis of HCHO. The agreement for simulations using pollutant mixtures containing only one organic species showed more variability. Formaldehyde predictions obtained for ethene, propene, and acetaldehyde were in excellent agreement under most conditions; significant differences, however, were noted for the 2-alkenes and aromatic hydrocarbons. Predictions of ozone and hydrogen peroxide obtained with the urban mixture were very sensitive to the initial concentration of HCHO included in the simulations. Depending on whether HCHO was assumed to constitute 0% or 2% of the initial mix of organic pollutants, resulting ozone yields varied by as much as 15--30%; hydrogen peroxide yields differed by factors of 4--10 depending on the mechanism used. To estimate the effectiveness of various control measures for limiting HCHO production, simulations were also conducted where individual organic species within the urban mix were reduced and the resulting decreases in HCHO yields were noted. The largest decreases in HCHO occurred for reductions in ethene and the other 1-alkenes; the smallest decreases were obtained when the monoalkylbenzenes or the alkanes were reduced.