LF bulbs are optimized for li-ion cells more than CR123 cells. The EO-M3T could be thought of as a li-ion compatible version of the MN16. (The MN16 is severely over driven on li-ion cells). What that means is that, the EO-M3T is also going to be under-driven on CR123s. The same could be said for the HO-M3T but to a lesser extent, which is just a hair below that point in power consumption.
I think the runtime figures on the LF website are quite optimistic for CR123s (I had never looked at the before personally). They probably calculated a runtime from cell capacity on paper minus an actual test. This type of runtime estimate works better with li-ion than with lithium primary cells beacuse li-ion cells hold closer to their rated capacity at a wider range of loads.
If I had to guess, I'd say that the HO-M3T would run closer to that 30 minute mark, and the EO-M3T would run about 20 minutes just like the MN16.
FYI: the MN16 and EO-M3T have about the same power consumption, but the MN16 will be brighter in this application because it has a lower design voltage that matches the discharge behavior of the CR123 cells better.
The HO-M3T might be worth a try. but I'm not sure if you will see a huge difference from the MN15 on CR123s. It will likely be slightly brighter, but possibly not by enough to notice. (probably less than a 35% difference)
As a very rough rule of thumb, to convert LumensFactory lumens into SureFire lumens for incandecent flashlights, divide LF by 2.
If the MN15 is rated 125 lumen, I would put the HO-M3T powered by CR123s at maybe 175 lumen.
Have you considered a pair of protected AW brand 17500 cells to power up the HO-M3T? That would give you 30 minutes of rechargeable runtime and a bit more impressive performance at this drain rate than CR123s, (more like "200" surefire lumens if I had to take a guess).
Eric