Patient breathing during cone-beam acquisition causes artifacts which can be reduced by taking into account the motion during the reconstruction of computed tomography (CT) images. Two main classes of methods are available. Gated or respiratory-correlated methods select a subset of the projections according to the respiratory signal. The selected projections are then used for image reconstruction. The dynamic reconstruction method compensates for the respiratory motion by using a 4D motion model which integrates all the projections. In this study, we compared the two methods with conventional analytic reconstruction using a phantom mounted on a mobile platform that translated sinusoidally in the cranio-caudal direction. The phantom was a wooden mass with a polyethylene cube in its middle. Projections were acquired with an Elekta Synergy system with the platform either fixed or mobile. We performed static reconstruction in both situations, and gated and dynamic reconstructions in the mobile situation only. The reference was static reconstruction with the fixed platform. Three criteria were used to quantitatively compare the reconstructions: blur criterion, signal-to-noise ratio and contrast-to-noise ratio. Results showed that the blur was highly reduced with gated and dynamic reconstructions. However, artifacts such as streaks and bands remained clearly visible in the gated CT image because only a subset of the set of cone-beam projections was used for reconstruction. Dynamic reconstruction significantly reduced these artifacts and yielded a quality close to the reference. This work demonstrates the potential of dynamic reconstruction to provide higher quality CT images than gated reconstruction when the 4D motion during the acquisition phase is known.