Last two years two different approaches were used for stock assessment of Antarctic toothfish in the Ross Sea. One of them, the CASAL model, (Dunn & Hanchet, 2007; Bull et al., 2007) is mostly based on likelihoods and potentially could insure proper mutual weighting of signals from all available sources of information incorporated into the model. The second one, the TISVPA model (Vasilyev, 2005, 2006; Vasilyev et al., 2006, 2007), takes care about robustness of analysis and includes a number of features aiming at consistent assessment using real (that is usually noisy and containing outliers) data. Besides that, some sorts of data, e.g. tagging data, are used in these models in quite different ways. Comparison of the results shows also that the input to the solution from different sources of information was also quite different: while in TISVPA all sources of data gave rather coherent signals about the stock (Vaslilyev et al., 2007), in CASAL the solution was mostly supported by signals from tagging data (Dunn & Hanchet, 2007), especially strong being the influence of fish tagged in 2006 and caught in 2007, driving the stock estimate down.