If it can run a small dynamo, it would be relatively simple to measure its power – with one caveat – using a resistor (e.g. a filament lamp), voltmeter and ammeter.
The caveat of course is this does not account for the energy expended within the dynamo, so it is testing the system, not the engine.
I was assuming the dynamo loss there would be quite small compared to the engine's rather obvious external-combustion flaw, so a full efficiency calculation – which would need you knowing the calorific value and quantity used, of the fuel – would still be a fair estimate.
I wonder if anyone has succeeded in making a brake-type dynamometer work in model scales?
(I recall a school physics experiment to measure the efficiency of a filament lamp "bulb". [Yes I know 'lamps glow, bulbs grow'.]. The lamp – a 6V one I hasten to add – was waterproofed and lowered into a small beaker of water containing a thermometer. I recall the electricity to light efficiency was well under 10%… >90% of the W being the heat.)