I (unfortunately) do not have a screenshot of me's reactor showing cop 10, but there was COP 1,49 and Barty reported 2,0 and 10 in the chat, this is of course not much reliable, but worth to be investigated!
I wouldn't doubt that there was an instant COP of 10 or even more.
It is not at all difficult to get a COP of 10 for a couple of seconds, and nothing extraordinary to be investigated.
Just visualize that the reactor/el.heater is operating with 1000W el. input power and the output also has stabilized at 1000W.
Then you drop (instantly) the el. input down to 100W. - Because of thermal inertia, the by the calorimeter reported output will still remain 1000W for some time - which will be calculated as a COP of 10 (but - of course - is total meaningless).