A couple of interesting notes about the 06Z runs, the GFDL has shifted back south of Galveston near the NHC track, and both the GFDL and HWRF show very little change in intensity (pressure-wise) until landfall. Of course, the max wind speed could increase even if the pressure doesn't fall, but the storm would have to contract its wind field to do that.

The ECMWF is the probably best model in the world. It is highly regarded by weather forecasters. Forecasters still usually prefer the model consensus at longer time ranges, though, which is why they were a little slow to initially bite off of the ECMWF solution for Ike.