The conversion is 3.414(kBTU/kWh). The COP*3.414 is called the heating seasonal performance factor. A COP=3.5 figure is 3.5*3.414=~12, about the same as a minisplit air source heat pump that costs a lot less.
Your question regarding the pump power is astute...it does reduce the effective COP significantly in geo systems (esp poorly installed one with too large pumps or deeper wells). In your case, it would reduce the COP by a factor ~5.1/(5.1+2.6) = 0.66, reducing the COP to ~2.3. This is too pessimistic since the air handler energy would appear as heat at COP = 1, I'd guess the COP is ~2.5.
In this case 100,000 BTUs would cost 100/(2.5*3.414) = 11.7 kWh.
My conventional cheapo air source HP (tricked out with defrost controls I hacked myself) is giving me COP=2.6 in a Philly climate, better than your quoted geo!
In general, in climates outside the northerm tier of the lower 48 and inland Canada, air source HPs provide a better ROI/lifetime operating cost than geos. Where are you?
check out:
http://homeenergypros.lbl.gov/forum/topics/geothermal-real-world-performance?xg_source=activity&id=6069565%3ATopic%3A105500&page=1#comments