Digital twin channel (DTC) is the real-time mapping of a wireless channel for the physical wireless communication in the digital world, which is expected to be one of the essential enablers for the sixth-generation (6G) communication systems and provides significant performance enhancements for air-interface design. In this work, we first define the evolution of five levels of channel twins with the progression of wireless communication. The autonomous DTC for 6G is elaborated from multi-dimensional factors such as methodology, characterization precision, and data category. Then, we provide detailed insights into the requirements and architecture of a complete DTC for 6G. Subsequently, a sensing-enhanced real-time channel prediction platform and experimental validations are exhibited. Finally, drawing from the vision of the 6G network, we explore potential applications and address open issues in DTC research.