Based on the understanding of inflation forecasts as an intermediate policy objective, this paper evaluates forecasts of different inflation models in Colombia during the inflation targeting (IT) period. The evaluation is done using three different statistical methodologies. The results suggest that the best models, in terms of precision, are the Food´s Relative Price and the traditional P* models. Additionally, a multiplier analysis is performed over these models, in which the sensitivity of short and medium term forecasts to shocks in the exogenous variables is assessed.