The 12z ensemble runs are out and up on Weathernerds.org now. These are often a better guide, albeit also quite early, for a feature that is not yet a TC and simply Invest tagged, than either the vort-centric models (best for bona fide TCs) or even the deterministic runs, that oversell what their particular biases are. E.g., significant development on virtually anything that spins by the operational GFS.

The early EURO ensembles (above) suggest that 95L would ultimately have a more northerly track if stronger, and a more westerly track if weaker, with a few exceptions.

The early GFS ensembles appear more evenly divided, with opportunities for both westward and northward tracks, regardless of intensity. This may be in part due to the early development bias still in these, along with an expectation for some higher net shear in the Caribbean, ending some runs and moderating others.