Objective: The purpose of this study was to evaluate the performances of the RHE/IL-18 assay using the SkinEthic™ RHE model for the identification of contact sensitizers. Methods: A set of 18 substances and mixtures was tested on this epidermal model, following the RHE/IL-18 protocol. The final results of the assay were obtained following 5 interpretation schemes, to determine the optimal prediction model for this assay with this specific test system. The data were analysed with a special focus on the basal level of IL-18 release and on the performance obtained with respect to three different gold standards: LLNA, HRIPT and an integrated reference, constructed from all available results. Results: No important differences were found in the performance levels depending on the three gold standards. The performances obtained with the SkinEthic™ RHE model support that this model may be considered as an alternative to different reconstructed epidermis models (EpiDERM™, EpiCS™ and VUMC-EE) for the performance of RHE/IL-18 assays. Conclusion: The prediction model to be used was refined, and more substances have to be tested in order to gather enough data for this evaluation and to determine the right criteria applicable for this assay using the SkinEthic™ RHE test system.