This paper critically evaluates the cracking model of MC 2010 with a view to the drafting of MC 2020. It emphasizes that there are differences in cracking behavior in tension and bending, which are not adequately recognized in the current model. Based on the analysis of experimental evidence, it shows that the current model has a significant skew in prediction performance when tension and flexure tests are considered separately. Based on these observations, modifications are proposed to account for the variation of tensile stresses