The different temperature datasets have differences in data sources and methods for filling in data gaps, and they often report their global mean changes with respect to different base periods. These differences in data/methods can lead to differences in global average temperature, but these differences are usually pretty small. To me, it looks like the global average temperature is pretty consistent among datasets, although there are small differences (like UAH had September 2024 as the warmest and ERA5 had it as the second warmest). Those differences look like they are within the typical range of uncertainty, but if I'm missing something, or if you noticed differences that look unusually large, please feel free to clarify.