Abstract The geometry of dust distribution within the inner regions of active galactic nuclei (AGN) is still a debated issue and relates directly to the AGN unified scheme. Traditionally, models discussed in the literature assume one of the two distinct dust distributions in what is believed to be a toroidal region around the supermassive black holes: a continuous distribution, customarily referred to as smooth, and a concentration of dust in clumps or clouds, referred to as clumpy. In this paper we perform a thorough comparison between two of the most popular models in the literature, namely the smooth models by Fritz et al. and the clumpy models by Nenkova et al., in their common parameter space. Particular attention is paid to the silicate features at ∼9.7 and ∼18 μm, the width of the infrared bump, the near-infrared index and the luminosity at 12.3 μm, all previously reported as possible diagnostic tools to distinguish between the two dust distributions. We find that, due to different dust chemical compositions used in the two models, the behaviour of the silicate features at 9.7 and 18 μm is quite distinct between the two models. The width of the infrared bump and the peak of the infrared emission can take comparable values; their distributions do, however, vary. The near-infrared index is also quite different, due partly to the primary sources adopted by the two models. Models with matched parameters do not produce similar spectral energy distributions (SEDs) and virtually no random parameter combinations can result in seemingly identical SEDs.