It looks like some models got the qpf right but the thermals wrong at 850. The lighter qpf models look like they're missing but were better on thermals.
It looked like Maine or somewhere here in 1996. There was snow piled up to the bottom of stop signs and in mountains in parking lots from Early January until around Valentines Day.
The warm nose that Holston mentioned in a prior post is often a problem for valley areas. It tends to be more prevalent towards Nashville and from Chattanooga to Knoxville. In this case there's also a general lack of heavy precip on some modeling and that doesn't cool the column as efficiently.
The Rufus is similar, and a bit south with the snow line, to the other hi-res models. The Rufus is supposed to replace the NAM and NAM 3k at some point.
The Rufus has 1.5 inches across downtown Knoxville.
The 03z RAP looks similar to the 21z with a decent strip of 2-3 inches of snow in the same areas.
The RAP and HRRR are more stark with rain or snow vs freezing rain or mix.
Euro is just a little too warm for most of Tennessee heading into the weekend, with very light precip. After, it gets cold, but barely any qpf, looks like warmth and moisture return together mid-month but that may be a head fake too.
The UKIE, GGEM, and ICON show a winter event of varying degrees in the area. The Canadian would be a major winter storm with significant ice along the 40 corridor from the mid-state to the mountains. Some snow on top of the ice across areas north of 40. UKIE looks like a long duration sleet event. The ICON is a 1-2 inch snow north of 40.