Sorry for the simplicity of the question, but I asked Google and she didn't know. Stackexchange is my next stop.
I took a brix reading of my steam beer wort at 13 brix (~1.052 SG), but it was probably about 110 degrees.
It's not usually necessary to cool - most refractometers have built in Auto Temperature Compensation (ATC) to correct for temperatures up to 30C/86F. However, for hot samples in a warm environment then some cooling may be needed.
When the ambient temp is cool (<20C/70F) the ATC will be sufficient and that the quantity of liquid is small and has a very small thermal mass compared to the body of the refratometer. So, it will cool quickly from mash temps, while only warming the refractometer by just a few degrees.
But if the ambient temp is close to the upper range of the ATC, then the additional sample may warm the body above what the ATC can handle. Then cooling is needed.
So, if in doubt, leave the refractometer for a few minutes. I've noticed a change of at least +0.5% brix after a couple of minutes when sampling hot mash samples, presumably because the body temperature rises above the level that ATC will correct.
Just called Robinar, manufacturer of refractometer I have. Posed the temperature compensation question to them and the response is: when calibrating the tool (IE: putting distilled water on the glass) make sure that test/calibtratiuon water is the same temp as the product/material to be tested.