Aim. To evaluate and compare the prognostic value of various scales on the risk of upper gastrointestinal bleeding (GIB) in patients with stable coronary artery disease (CAD) in the Russian population.Material and methods. The study included patients with stable CAD — participants of the REGATTA-1 prospective single-center registry. The number of points was assessed according to the reference score of the 2015 European Society of Cardiology (ESC), PRECISE-DAPT, ABC-HBR and REACH scores, as well as two Russian scores — the REGATTA score and the ORACUL score, originally developed for patients with acute coronary syndrome (ACS).Results. A total of 623 patients were included (median age 60 [53, 68] years; men — 78,7%). Over 2,5 [1,1-14,7]-year follow-up, the incidence of major and clinically significant GIBs was 1,9 cases/100 people/year. There were following score with sufficient predictive value: REGATTA — Area Under the Receiver Operating Characteristic Curve (ROC AUC) 0,865, p<0,0001, F1 measure 0,4; ESC 2015 — ROC AUC 0,791, p<0,0001, F1 measure 0,243; PRECISE-DAPT — ROC AUC 0,713, p=0,001, F1 measure 0,397, and ABC HBR — ROC AUC 0,743, p<0,01, F1 measure 0,432. The REGATTA score is characterized by the greatest predictive significance (p<0,05 for all pairwise comparisons). The original ORACUL scale (ROC AUC 0,507, p=0,9, F1 measure 0,150) and the REACH scale (ROC AUC 0,572, p=0,164, F1 0,177) did not significantly predict the GIB risk. The identified patterns maintained when analyzed with binary values "high risk/no high risk" for each of the scales. Modified ORACUL scale for stable coronary artery disease after recalibration using binary analysis was not inferior to the ESC 2015, PRECISE-DAPT and ABC-HBR scales (ROC AUC — 0,645 (p=0,002)).Conclusion. The universal PRECISE-DAPT and ABC-HBR scales can be used to assess the GIB risk of along with the 2015 ESC score in patients with stable coronary artery disease. The Russian REGATTA score has the best predictive ability, but its limitation is the lack of external validation. Modification and recalibration of the ORACUL scale significantly increases its predictive value in patients with stable coronary artery disease.