Clearly, it did get hot enough to collapse, because it collapsed.
The NIST modelling showed temperatures high enough to cause collapse, and the distribution of temperatures in the model agreed with the identifiable physical evidence. That is, the model equivalents of the core columns that actually reached 250, and so did not have identification numbers burned off, also only reached 250. Other locations in the model reached much higher temperatures, for significant periods of time. No doubt there were sections of core column that didn't get much above room temperature before the collapse.
The fire modeling, then, was consistent with the building failing. To make the argument that the fires did not heat the steel enough is to argue that the fire modeling is wrong.
I have seen criticisms of the NIST fire modeling methodology from other organizations with expertise in fire modeling, but the criticisms that I have seen I interpreted as NIST being too conservative with the fuel loading in the models. Some seem to think that even had all the fireproofing stayed attached, the buildings would still have failed. NIST, as I understand it, concluded that it was a combination of removal of fireproofing and widespread fire that led to the collapse.
So, NIST conservatively models it as hot enough to fail the members that they think failed. Experts outside of NIST review the methodology, and complain not about overestimation of the heating of the steel in the hottest areas, but underestimating it.
Do you have some particular reason why you think the NIST modeling overestimated peak temperatures, and why other experts didn't catch it, instead arguing that the temperatures could have been even higher?