So, the 00Z NAM is much snowier than the previous couple of runs of the NAM for 95 and even for the coast. Doesn't mean it's right, but by definition, models should be experiencing less and less errors in track, precip type and intensity, as one gets closer to the event, such that a model should almost perfectly predict 1 minute ahead and very accurately predict 1 hour ahead and reasonably accurately predict 6 hours ahead.
Since the NAM is the first model to run with 7 pm EST data, which ingested extra hurricane hunter flight data, I'd expect there's a good possibility the other models will respond similarly when their 00Z output comes out over the next 3 hours (some take longer to run than others, but they all start with the same data), but that's not a given.
We're now only 12 hours from the most critical part of the forecast, involving the formation and movement of the surface low pressure, so one would hope all of the models will be getting closer to the actual outcome, but unfortunately, in very dynamic, complex scenarios, like this, errors are still present only a few hours out.