Since 2012 The Bank of Russia has been pursuing an active policy of implementing an approach to calculating credit risks that is based on internal bank ratings in accordance with the document “International Convergence of Capital Measurement and Capital Standards: New Approaches” [1] of Basel Committee on Banking Supervision. This approach is called IRB approach. Qualitative system of internal ratings facilitates an effective solution of the following tasks by the bank: to increase the quality of lendees assessment and lending-related decision-making, to determine lending limits, to perform pricing with account of risks, to evaluate a credit risk premium (spread), to substantiate the level of reserves against possible bad debts, to calculate RWA and sufficiency of the regulatory capital, to calculate the amount of the economic capital, RAROC, to improve the processes of risk and data quality management. Therefore, notwithstanding the plans of transition to IRB-approach for assessment of regulatory capital, it is relevant for banks to develop and improve own rating systems. In this article, we will consider a number of practical aspects of IRB-models creation that are related to determination of the parameters of discretisation and dynamic conversion of factors, the use of macroeconomic variables as factors and mapping of a model with an international scale.