AIMS: Dynamic risk models update the risk profile of ST-elevation myocardial infarction (STEMI) patients over the acute period following the event and have implications to clinical practice and research. METHODS AND RESULTS: Multivariable survival models were developed in 5,745 STEMI patients undergoing primary percutaneous coronary intervention (PCI) enrolled in the APEX-AMI trial to predict 90-day mortality from 4 clinically relevant times: baseline, 2 hours, 24 hours, and 96 hours. Culprit coronary thrombolysis in myocardial infarction flow grade, 30-minute post-PCI worst-lead ST-elevation residual, and in-hospital clinical events were considered in the models. The 90-day mortality was 4.7%; the cumulative proportion of mortality occurring within 2, 24, and 96 hours was 8%, 22%, and 40% respectively. Relative to the baseline risk factors, age and systolic blood pressure remained highly ranked in the post-baseline models. However, the relative importance of heart rate, Killip class, and creatinine declined, whereas markers of coronary reperfusion and in-hospital events (shock, congestive heart failure) became increasingly influential. The c-index increased from 0.819 at baseline to 0.847 at 96 hours. Over the forecasting periods, the proportion of “low-risk” (<1.1% 90-day mortality) patients increased from 20% to 49%. This approach derived from an unfolding series of models reveals the shifting levels of mortality risk from baseline to 96 hours. CONCLUSION: This novel approach in STEMI patients undergoing primary PCI demonstrates the dynamic nature of risk over time and may prove useful in understanding risk and in clinical decision making.