Training models

in week 3, first lab there is classification model. when getting the results it says that Model 1 is the smallest and Model 3 is the largest. therefore model 1 is the best approach, but Model 1: Training Set Classification Error: 0.05833, CV Set Classification Error: 0.17500
Model 2: Training Set Classification Error: 0.06667, CV Set Classification Error: 0.15000
Model 3: Training Set Classification Error: 0.05000, CV Set Classification Error: 0.15000 . im seeing that model 3 is the lowest number am I wrong?

We don’t care about the classification error on the training set.
We care about the CV (or Test) set results.