Abstract: Aiming at the lack of standardized evaluation metrics for advanced capabilities in AI voice interaction within smart cockpits in the era of large language models, this paper constructs a ...