Under ultra-high vacuum (UHV) conditions, CO oxidation on Pt occurs via the three-step Langmuir-Hinshelwood mechanism including CO adsorption, dissociative oxygen adsorption and reaction between adsorbed CO and O particles resulting in formation of CO 2. Assuming that this mechanism is operating also at high (atmospheric) pressures, we have explored the steady-state reaction kinetics predicted in a wide pressure range by several kinetic models. Focusing on the pressure-gap problem, we have arrived at the following conclusions. (i) If one in simulations employs the kinetic parameters deduced from UHV experiments but ignores the coverage dependence of the CO desorption and CO + O reaction rate constants, the Langmuir-Hinshelwood scheme fails to describe the apparent activation energy and reaction order with respect to CO at high pressures. In addition, the predicted values of the ratio P CO/ P O 2 corresponding to the ignition locus on the stability diagram are in poor agreement with the experiment at high pressures. (ii) The mean-field corrections, taking into account a weak coverage dependence of the rate constant for CO desorption (e.g., UHV data for CO desorption on Pt(111)), are not a sufficient modification to properly describe the reaction kinetics at high pressures. (iii) If one, however, explicitly introduces into the model a strong repulsive nearest-neigbour CO-CO lateral interaction, the predicted results are in reasonable agreement with experiment, both at low and high pressures. In particular, the calculated apparent activation energy and reaction orders are in the same range as the measured ones. The ignition locus on the stability diagram is also predicted at reasonable ratios of P CO/ P O 2.