We investigated the ability of a reduced chemical kinetic model of 18 reactions and 13 active species to predict the heat release for a blend of primary reference fuels with octane rating 63 in a motored research engine. Given the initial fuel-air mixture concentration and temperature, the chemical kinetic model is used to predict temperature, heat release and species concentrations as a function of time or crank angle by integrating the coupled rate and energy equations. For comparison, we independently calculated heat release from measured pressure data using a standard thermodynamic model. We found that: i) the induction time of heat release can be matched when the rate parameters of the RO2· isomerization reaction are chosen between the values suggested for iso-octane and n-heptane; ii) the rate of heat release predicted by the kinetic model is much greater than that calculated from the experiment; and iii) the kinetic model underpredicted the specific heat release by more than 50%. Analysis of the chemical kinetic model showed that in its present form, the specific heat release could not be matched. The original chemical kinetic model was extended by adding 11 reactions and 5 active species to account for the oxidation of aldehydes, olefins, carbonyls and the formation of CO. This improved the match of both the specific heat release and the overall heat release profile, as well as providing new capabilities for predicting carbon monoxide production which is a key indicator of preignition chemical activity.