What is the impact of overfitting in a machine$-$learning model?

Question $2$ options:

result in high accuracy for the training dataset, but poor performance for the testing dataset and poor model generalization

result in high accuracy for the training dataset, high performance for the testing dataset and poor model generalization

result in low accuracy for the training dataset, poor performance for the testing dataset and poor model generalization

result in high accuracy for the training dataset, but poor performance for the testing dataset and high model generalization

result in high accuracy for the training dataset, high performance for the testing dataset and high model generalization