Abstract: This study constructed a cost-benefit meta-assessment model based on the reflection of existing meta-assessment models and determined the gray correlation between the cost of performance management and the completion rate of performance management goals in Guangxi’s municipalities. The results of the study show that performance management in Guangxi, which has had high efficiency for 8 consecutive years; the most efficient of all performance management costs is livelihood assessment input, followed by key work input, funding input and manpower input. Meanwhile, the accuracy and robustness of the cost-benefit model were verified by using the comparative analysis of k-means clustering results and fixed-effects model, and then the effectiveness of Guangxi government performance management was systematically verified. The cost-benefit model provides a set of scientific assessment methods and index systems for performance management meta-assessment of municipal governments, thus forming a new meta-assessment perspective and idea, and providing empirical evidence for optimizing administrative resource allocation.