Motors don't work the same so there's no real need to run at higher RPM. What DOES happen is that there's going to be some point where the internal DC resistance becomes significant at higher current levels and it drops the overall efficiency. What you want to do is pick a motor that you can run at the best peak efficiency for current and voltage. That'll give you the most output per lb of motor and battery weight. The RPM where this will occur is set by the Kv value. I've not seen any real advantage of high Kv motors over lower Kv motors other than when wound on the same frame the higher Kv value options use less turns of heavier wire than the low Kv option. So the DC resistance is lower for the high Kv version which means there's less heat generated at higher currents.
As for gears they will always soak up SOME power both for the friction of the teeth meshing as well as the additional bearings. So if they can be avoided then it is the better option. What determines if you go geared or not is how the propeller efficiency climbs with an increase in size vs the amount of power used up by any reduction gearing. To keep things simple I would strongly suggest sticking with direct drive low Kv motors at first.
The other good news is that since you'll be wanting the prop and motor to work best at a relatively low climb speed you can perform ground based static thrust testing for the props and actually trust the data to a much higher degree than for normal model flying. The options you'll find work the best will be the ones which provide the most thrust and those which are a tad higher in pitch and diameter without running the current consumption up higher than you want. The reason I say you can use the ones that are a LITTLE higher in pitch and diameter is because the drop off with pitches higher than around 8 inches occurs due to the blades stalling during static thrust testing. But once in the air for prop testing you may well find that those "nearly" options work better than the best ground performers.
Some of the older F3J class gliders were fine with polyhedral and rudder plus elevator. But in looking up images of those from today they seem to have downsized a little and gone to ailerons. But there were a lot of 4'ish meter span models that flew back 10 to 15 years ago with rudder, elevator and flaps or spoilers. And it's not like you'll be doing aerobatics with it...
To get the most out of dihedral in this way it's best to go with a 5 to 7 panel style wing with each panel towards the tips becomeing smaller and at a higher angle to simulate the shape of an elliptical curve as much as practical. It puts the strongest responding part of the dihedral out at the tips where it has the most leverage to roll the model from the yaw to roll coupling.