I don't have M61WLL to test - but I do have M60LL so this shall suffice for the time being, until I get M61LL. I assume their behavior is pretty similar, with M60LL having marginally higher drain. You can take this info and extrapolate the behavior of M61LL and be pretty much on the money.
I ran M60LL on 1x123 Panasonic cell, 1550 mAh capacity, in a FiveMega 3P clone body. That is one small device and an excellent candidate for EDC.
Immediately I got impressive brightness, almost full lumens you can get from the module -- for what seemed like about 45 minutes to 1 hour. Then it drastically dropped. I checked the cell and could not believe it went from full charge of 3.3V to 2.9V in under 1 hour. I thought I would get much better runtime than that. The lumens dropped to roughly in half within that time period. I was getting worried.
I assumed it would continue to decline at the same rate. However it's usage was not linear at all, and it stuck at 2.9V for a very long time, for the next 4 hours. The lumen level I would estimate at somewhere between 30-40 lumens. Decent but obviously lesser than than of on 1x18650 and surely lesser than multiple cell configurations.
Then I left the light turned on overnight and 6 hours later, it has diminished a little more but still very obviously bright and useful. Much brighter than the 10 lumens of Gerber Infinity Ultra I was checking against. So the runtime was probably close to 10 hours and it's not even done yet, seems to be halfway done. I will resume checking tonight.
Behavior should be even more impressive with M61LL, due to its lower drain.
As one data point, M60LL runs for 90 hours on 3xAA Alkaline cells, 40 hours on 2xAA Akalines and should run obviously less than that on 1x123 due to reduced capacity, I am estimating somewhere between 12-20 hours.
Surefire 3P clone by FiveMega + MxxLL is an excellent candidate for TSHTF outage and certainly EDC.