Looking at the last 4 runs of the GFS, only on the 12z run (3rd run out of 4) did the GFS pick up on the lake effect snows. By then, it was approximately noon CDT, and got a solid handle on it at about the same time.
Looking at this, the GFS obviously didn't perform well.
The NAM got a solid handle on the snows at 12z as well, and the NAM latched onto GFS thinking that the lake effect snows would hit Chicago at the 12z run.
The SREF, which typically is the best in short-hand situations, always had a thinking that lake effect snows would occur in SE Wisconsin, but never got the feeling for it hitting Chicago until it was happening.
In conclusion, the models did not handle this well at all, and the rating I would give the models a rating of a D+.