Without knowing the fitness function you’re using or the problem you’re trying to solve or the settings you chose, there’s not much I can say about this value, except that it’s some best fitness value that you achieved in a run. However, all the built-in fitness functions of GeneXproTools are normalized between 0-1000, so this gives you a good idea of where your model stands in terms of performance. Notwithstanding, it’s always good practice to use a more standard metric of model performance (such as the r-square or correlation coefficient for regression models or the classification accuracy for classification and logistic regression models) in order to make comparisons between different models, especially if they are created using different fitness functions.