The choice of a dependence structure between default times drives the prices of basket default swaps and CDO tranches. We therefore assess the model risk associated with the pricing of multiname credit derivatives. We discuss the comparison methodology and consequently we consider different pricing models associated with different copulas of default times: Gaussian, Student t, Clayton, Marshall-Olkin, double t. We emphasize the use of stochastic orders to derive some properties of CDO tranche premiums. It can be shown that base correlation tranches premiums increase with some dependence parameters. We also compare semi-explicit pricing approaches and the use of large portfolio approximation techniques.