Vortex strength is directly linked to Cl. A model in level flight will produce a much weaker vortex pair than a model in a high-g maneuver. A model flying a ballistic trajectory (think Vomit Comet) produces no vortex.

A model in level flight at high speed will produce a smaller vortex than the same model in level flight at low speed.

Think of it this way: While a model is in flight, the wing is producing a region of higher pressure below the wing and lower pressure above the wing. Since the air to the side of the wing is at some intermediate pressure, the air below the wing will have a tendency to gain an outward velocity component and the air above will tend to gain an inward velocity component as the air tries to return to a uniform pressure. Since the pressure gradient across the wing is the same regardless of speed (we're assuming zero washout for simplicity) the outward and inward acceleration of the airmass is the same for high and low speeds. At high speed, the air spends less time under and over the wing than it does at low speeds. Less time accelerating = lower final velocity. Therefore, a fast plane produces a weaker vortex than a slow plane.

End plates act as a barrier to this inward and outward acceleration, and in doing so, reduce the strength of the resulting vortex.

Winglets act partially as end plates, and also as a vertically-oriented high-washout wingtip. The angle of attack of the winglets is usually set to produce a vortex in the opposite direction of that naturally produced by the wing, thereby partially cancelling the lift-induced vortex.

Both of these approaches effectively increase the aspect ratio of a finite wing. The exact design intricacies make it difficult to design a truly effective wingtip device, and a poorly designed device will actually reduce performance. So, for model purposes, they're probably more trouble than they're worth (unless you have access to a powerful suite of CFD software).