I think the GFS 6z shows that we shouldn't write off chances of breaking records just yet.
We have seen this before. In August 2003, the peak of the heat was forecast to be gone before 9th / 10th August. However, as it approached, the models trended hotter, which is what actually transpired.