The Year Of Polar Prediction (YOPP) dataset of the European Centre for Medium-Range Weather Forecasts (ECMWF) contains initial condition and forecast model output from the operational global, coupled numerical weather prediction system. The dataset has been created to support model forecast evaluation, predictability studies and model error analyses over polar areas, which are strongly affected by climate change with yet unknown feedbacks on global circulation. The dataset complements YOPP observation and modeling research activities that represent a key deliverable of the World Meteorological Organization's Polar Prediction Program. The dataset covers the period from mid-2017 until the end of the MOSAiC field campaign, expected for autumn 2020. Initial conditions and forecasts up to day-15 are included for the atmosphere and land surface for the entire period, and for ocean and sea-ice model components after June 2019. In addition, tendencies from model dynamics and individual physical processes are included for the first two forecast days. These are essential for characterizing the contribution of individual processes to model state evolution and, hence, for diagnosing sources of model error.