GoCAV is an entirely in-house developed autonomous platform, designed for real-time control and decision-making in complex environments. In the latest track session, 10 diverse real-world scenarios provided critical insights that enhanced the performance of the steering neural network. Thanks to clear weather and a controlled testing environment, the team achieved consistent and focused model training. GoCAV showcased advanced long-range object detection, accurately identifying obstacles beyond the immediate roadside through integrated vision and perception systems. During each run, all vehicle parameters were continuously logged via real-time telemetry, enabling in-depth analysis and system refinement. Every test brings GoCAV closer to achieving safe, scalable, and mission-ready autonomy.