I do tend to rely heavily on the default favored ECMWF model for my cloud coverage prediction needs for airborne survey priority planning, but then temper its depiction with the ICON and the HRRR models. When those three are in decent agreement then I take that as a high confidence level. When they are disparate then I drill down more in to other atmospheric feeds (for example the realtime water vapor depictions) and look at what NAM & GFS have to say.