We document that reliance on internal ratings-based (IRB) models to compute credit risk and capital requirements reduces bank opacity. Greater reliance on IRB models is associated with lower absolute forecast error and reduced disagreement among analysts regarding expected bank earnings per share. These results are stronger for banks that apply internal ratings to the most opaque loans and adopt the advanced version of IRB models, which entail a more granular risk assessment and greater disclosure of risk parameters. The results stem from the higher earnings informativeness and the more comprehensive disclosure of credit risk in banks adopting internal ratings. We employ an instrumental variables approach to validate our findings.