In this paper, Bayesian accelerated life testing models that make use of the generalised Eyring time transformation function are compared. In accelerated life testing, products are tested under more severe than their normal operating conditions. This is done to obtain failure data in a much shorter timeframe, where the life characteristics of the product under the normal operating conditions can then be extrapolated. The models compared incorporate two stress variables, one thermal and one non-thermal. The Weibull, Birnbaum-Saunders and log-normal distributions are used as the life distributions. The models are applied to a real dataset, where various prior settings are considered as part of a sensitivity analysis. Due to the intractable nature of the posterior distributions for these models, Markov chain Monte Carlo methods are employed to generate posterior samples for inference. The deviance information criterion is used to compare the fit of the models and the predictive reliability is also calculated.