Evaluating Vision-Language-Action Models in Robotic Manipulation: Performance, Implementation, and Comparison with Deterministic Systems