In the Internet of Vehicles (IoV), various latency-critical and data-intensive applications have recently emerged to support smart traffic solutions. The sixth generation mobile networks (6G) greatly reduce the communication delay for the latency-critical tasks. However, the computing resources and execution efficiency for data-intensive tasks are still inadequate. The stationary edge computing (EC) servers, on the other hand, lack the flexibility to give service to moving vehicles. Therefore, Cybertwin is introduced in the IoV paradigm to provide a unified access point for EC. In addition, the 6G-enabled network in box (NIB) is deployed in vehicles to provide flexible computing power. However, in this article, the optimization of NIB task migration is still a challenge; thus, NIB task migration method (NTM) for IoV is proposed. The Pareto envelope-based selection algorithm is employed to determine the strategy. Finally, NTM is evaluated by a real-world dataset of the service requests.