I've become very suspicious of any even medium-range GFS forecasts. There are a number of reasons for this that I won't go into here, but long and short of it, generally speaking, the GFS has become one of the last models I look to for anything more than a very broad-brush of future conditions. That said, last week's 5/27/0z run out to today arguably performed better than its competitors. While a closed circulation still looks to be a good bit more elevated than at the 925mb level predicted, it wasn't a totally outlandish modeled outcome, and as mentioned, seemed to outperform the EURO, EURO AI, GFS AI and CANADIAN runs from the same time.

Below: GFS 5/27/0z for 6/4/12z

Below: Actual 6/4/18z

Something I'll be keeping in mind as the current season continues.