The Imperial model had larger errors, about 5-fold higher than other models by six weeks. This appears to be largely driven by the aforementioned tendency to overestimate mortality. At twelve weeks, MAPE values were lowest for the IHME-MS-SEIR (23.7%) model, while the Imperial model had the most elevated MAPE (98.8%). Predictive performance between models was generally similar for median absolute errors (MAEs)
https://www.medrxiv.org/content/10.1101/2020.07.13.20151233v5.full