Multivariate geometric risk measures are tools allowing for flexibility in the choice of the confidence levels; each random variable can be studied at a specific - possibly different - levels. The resulting vector from risk geometric measures has the same dimension as the number of random variable, and considers the global aggregate portfolio as well as individual protection levels. Properties of geometric VaR, TVaR and RVaR will be provided. Illustrations and comparison with other multivariate risk measures will presented.