The consumption of energy by Machine Learning (ML) has increased significantly. There is growing concern about the sustainable use of ML, where choosing the best ML model should also consider energy efficiency. The main objective of the Green Machine Learning paradigm is the simultaneous optimisation of accuracy and energy consumption. The literature has presented some suggestions for metrics to be used. However, these metrics have not been extensively compared among different ML models. To address this aspect, in this paper, we have analysed six Machine Learning models applied to three benchmark datasets for binary classification tasks, focusing on performance and energy consumption. The results of the F1- Score show that the random forests model outperformed the other models, while logistic regression was more energy efficient. These results demonstrate the trade-offs between model performance and energy consumption, providing valuable guidance for algorithm selection. Performance metrics are an essential benchmark, with Python’s Scikit-Learn suite of models often outperforming neural networks in classification tasks. Future research should extend energy analysis to other machine learning methods and consider metrics that balance performance and energy consumption.