Small scale probes implementing shape memory alloy (SMA) actuation show great promise in applications requiring remote and minimally invasive access to small environments. Such environments include physiological spaces like those located in human and animal bodies as well as cavities within mechanical systems. Probes examined here are generally snake like in appearance composed of one or multiple independent segments, which in turn are made up of one or multiple SMA actuators performing work against an elastic spine. As the actuator(s) of a given segment are activated, the spine bends causing the probe to bend in the area of that segment. When the actuator(s) are deactivated, the force generated in the bending of the spine returns the segment to its neutral position. Activation and deactivation of actuators is accomplished by heating and cooling respectively, enacting the solid phase changes that are characteristic to the shape memory effect. The gage of control over probe shape depends on the number of independent segments that are available per unit length and the degree of control an operator has over each of the segments. The work presented here discusses the constraints imposed on the design of SMA actuated probes, and how those constraints become more critical and limiting with reduced physical scale and refinement of motion control. Numerical and finite element models have been developed showing the interrelationship between mechanical design, the thermal and phase states of the SMA actuator(s), and the mechanical performance of the total system. Performance concerns examined include probe shape control and the limits of shape change as a function of physical scale. Comparative data is presented between behavior predicted by the models developed and performance observed during the testing of prototypes. It is concluded that segment length, linked to refinement of probe control, is limited by its thermal boundary conditions.