This is paper presents a state-of-the-art gas-phase mechanism for the degradation of α-pinene by OH and its validation by box model simulations of laboratory measurements. It is based on the near-explicit mechanisms for the oxidation of α-pinene and pinonaldehyde by OH proposed by Peeters and co-workers. The extensive set of α-pinene photooxidation experiments performed in presence as well as in absence of NO by Nozière et al. (1999a) is used to test the mechanism. The comparison of the calculated vs measured concentrations as a function of time shows that the levels of OH, NO, NO2 and light are well reproduced in the model. Noting the large scatter in the experimental results as well as the difficulty to retrieve true product yields from concentrations data, a methodology is proposed for comparing the model and the data. The model succeeds in reproducing the average apparent yields of pinonaldehyde, acetone, total nitrates and total PANs in the experiments performed in presence of NO. In absence of NO, pinonaldehyde is fairly well reproduced, but acetone is largely underestimated. The dependence of the product yields on the concentration of NO and α-pinene is investigated, with a special attention on the influence of the multiple competitions of reactions affecting the peroxy radicals in the mechanism. We show that the main oxidation channels differ largely according to photochemical conditions. E.g. the pinonaldehyde yield is estimated to be about 10% in the remote atmosphere and up to 60% in very polluted areas. We stress the need for additional theoretical/laboratory work to unravel the chemistry of the primary products as well as the ozonolysis and nitrate-initiated oxidation of α-pinene.