In the modern digital society, the deep learning models are used by researchers for detection of human breast cancer. Those models are initially tested on various standard datasets. The performance of those detection models is measured by the test results and pre-defined metrics. On the other hand, Multi-criteria Decision-Making (MCDM) algorithms are also used together to enhance the performance of breast cancer detection methods by effective decision-making. To find the performance of significant deep learning models for efficient detection of breast cancer, benchmarking methods are executed by the researchers on standard datasets. As the benchmarking methods generate data of various metrics, it becomes difficult to find out suitable deep learning model for this purpose. This study introduces a complete methodology for post-benchmarking, specifically developed to evaluate and rank four significant deep learning models used in the detection of breast cancer. By utilizing a comprehensive benchmark dataset, we systematically assess the performance of the models through the application of well-established measures. In order to optimize the decision-making process, two different MCDM methodologies are utilized together, namely, the first one is Analytic Hierarchy Process (AHP) and the other one is Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). Findings of our study reveal the advantages and disadvantages of each model, offering a great resource for professionals and scholars in the field to make informed decisions when choosing the most effective deep learning approaches for detection of breast cancer.