Introduction: In circumstances where the ante mortem list is unknown, gender determination would exclude onehalf of the population, aid in a more precise search of the ante mortem records. This study aims to formulate gender prediction models in the Pakhtun Pakistani population using digital dental arch dimensions. Methods: Data collection and analysis of the dental casts were conducted on 128 subjects, 64 males and 64 females from the Pakistani population. The mean age of the subjects was 19.2 years old. Several linear dental arch dimensions were measured and recorded for both upper and lower arches. Results: It was found that gender differences in linear arch dimensions were statistically significant for both males and females (p<0.05); in which the arch dimensions for the males were larger than the arch dimensions for the females. Stepwise discriminant function analysis found that the highest discriminant power of the variables was present within the inter-second premolar width for the upper arch and inter-molar width for the lower arch. These variables significantly contributed to gender variance. Moreover, the prediction of 67.2% of original grouped cases for the upper arch and 66.4% of cross-validated group cases was correct. Similarly, the correct prediction was made on 64.8% of original grouped cases for the lower arch and 64.1% of cross-validated group cases. Conclusion: The dental arch dimensions were larger among the males compared to the females. Prediction models obtained in this study were moderately strong predictors which may be used as an adjunct to predict gender.