Mutation validation as a complement to existing applied machine learning validation schemes hasbeen explored in recent times. Exploratory work for Learning vector quantization (LVQ) based onthis model-validation scheme remains to be discovered. This paper proposes mutation validation as an extension to existing cross-validation and holdout schemes for Generalized LVQ and its advanced variants. The mutation validation scheme provides a responsive, interpretable, intuitive and easily comprehensible score that complements existing validation schemes employed in the performance evaluation of the prototype-based LVQ family of classification algorithms. This paper establishes a relation between the mutation validation scheme and the goodness of fit evaluation for four LVQ models: Generalized LVQ, Generalized Matrix LVQ, Generalized Tangent LVQ and Robust Soft LVQ models. Numerical evaluation regarding these models complexity and effects on test outcomes,pitches mutation validation scheme above cross-validation and holdout schemes.