After a run, the fitness and R-square/Accuracy for the best-of-run model are shown both for the training and testing sets. For the remaining models only the fitness and R-square/Accuracy obtained during training are shown, but you can easily access the fitness and R-square/Accuracy for the testing set by clicking the Test All or Test Current buttons.

Each one of these intermediate models can be activated by clicking on it. And once a model is active, it can be used to make exactly the same kind of analysis and transformations one does with the best-of-run model, that is, look at its structure (size, shape, and elements), translate it into different programming languages, score a database, explore a different evolutionary path by using it as seed, check its statistics and how it generalizes in the Results Panel, twinkle with its structure in the Change Seed window, evaluate its performance in a different testing set, and so forth.

For your reference, in the History of the Report Panel, the active model is highlighted in yellow.