Interesting question, and I can only answer in general terms because much depends on speculation.
Short answer:
- use one central exhaust
- one big blower is better than two small ones
- one short flow channel is better than two long ones
- suction on top might be beneficial for high lift
Radial-flow compressors are very light and achieve high pressure gains per stage. However, the simple plastic thing in your illustration will not have good efficiency. Generally, cambered blades are better than straight blades, and the thinner they are the better. Thick blades block more of the flow path. The drawback is their high diameter, and by placing the blower on its side, you get around this obstacle in an elegant way.
The idea of using two exhausts complicates things; if you use a single one, you can avoid problems due to unsymmetrical thrust. Given that the compressor runs in one of two directions, asymmetries will be very hard to avoid. By placing vanes into the exhaust you can actually give the model plane some degree of sideways thrust vectoring. Also, two flow channels will have more surface area per cross section than a single channel. Your losses will be higher if you use two like in the illustration because they are longer than necessary.
The same goes for the blower: A single, larger one will always be more efficient. The gap size in a small blower is probably relatively bigger than that of a larger blower, and gaps are real efficiency killers.
Another efficiency killer is flow separation in the ducts. Make sure that cross section changes are gradual and avoid bending the flow as much as possible. It will probably help if you install cambered guide vanes at the entry of the exhaust to convince the rotational flow to change direction and leave the round compressor chamber.
Suction on top is also interesting. There have been many ideas about boundary layer suction to increase maximum lift, and your turbine inlet will certainly help to remove some of the boundary layer where it will separate first. Normally, turbine intakes are placed where the normal flow field already shows some pressure increase due to ram effects, but at the speed of a model airplane this will not be a factor. I expect the best intake in your case is a hole on top of the aircraft with nicely rounded edges. Those rounded edges should already produce some lift, and the whole configuration will probably achieve a higher stall angle of attack when the engine runs at full power than at idle power.
I can only speculate how much power will be converted to thrust, but I bet placing an ordinary propeller at the usual place will be more efficient. But the configuration you propose has its own merits.