This dataset presents a comprehensive probabilistic analysis comparing five existing compressive strength-porosity models for OPC composites, shedding light on the uncertainties associated with compressive strength predictions. Traditional models for ordinary Portland cement cannot often quantify uncertainty due to factors like diverse testing protocols and limited data. This study follows Bayesian Analysis Reporting Guidelines (BARG) and employs a 95% High Posterior Density (HPD) Interval and the Range of Practical Equivalence (ROPE) as combined decision criteria. The comparative evaluation reveals the probabilistic version of Ryshkewitch's model as the most plausible, providing valuable insights for enhancing the reliability of compressive strength predictions. This dataset contains extracted data and results from the analysis, contributing to a robust understanding of the models and their implications in risk analysis and decision-making for cement-based materials.