The GFS does not handle the cold air well on the 12z run...NAM seems to handle it the best.
The GFS initialized (12z) with 59 degrees in Wichita, when in reality it was 63 degrees. At 15z, the GFS had it at 37 degrees, when in reality it was 34. By 18z, 30 mins from now, the GFS predicts a temperature of 44 degrees.
Wichita is currently at 31.
The 12z NAM initialized much better. At 12z, it initialized with 61 degrees. At 15z, it had 32. It is predicting an 18z temp of 32 degrees for the top of the hour.
Probably best to follow short range meso models for tracking the cold front and temps. As for precip, it would appear the NAM has been most consistent in that area regarding big ZR totals, but there are questions remaining on how shallow the cold air will be and if the precip falls heavy enough that it doesn't accumulate and warms the column from the top-down.