Human reconstructed skin models could be very useful tools to quantify percutaneous permeation and absorption. Before using such models, the reproducibility in the same batch and/or various batches, i.e. the relevance of the results obtained, must be verified. The reproducibility of 3 industrial models--EpiDerm, Episkin and SkinEthic--was tested regarding the permeation and skin absorption of 3 topically applied compounds (with a large range of physicochemical properties): lauric acid, caffeine and mannitol. For all the models, the intrabatch reproducibility was greater than the interbatch reproducibility. According to the batches tested, the larger difference in terms of reproducibility between the 3 models was observed in the case of mannitol, a very poor permeant. In this case, the best reproducibility was observed with EpiDerm and Episkin. Moreover, the rank order of the 3 compounds applied, in terms of permeation and skin absorption, was the same as that expected from ex vivo human skin. Such results revealed human skin models as a promising means to test in vitro permeation and percutaneous absorption of topical products.