Accurate, high spatiotemporal resolution precipitation data are essential for effectively monitoring, forecasting, and mitigating natural hazards. The Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement (IMERG) Version 07 (V07), released in July 2023, has drawn significant attention, due to the robust performance of its predecessor (V06). However, a comprehensive regional-scale evaluation of the IMERG V07 datasets with varying latency (Early, Late, and Final runs) and their improvements over V06 remains lacking. This study aims to systematically evaluate and compare the performance of IMERG V07 and V06 over the Yellow River Basin (YRB) and its sub-regions, using gauge observations as reference. This evaluation focuses on precipitation estimation accuracy, rain/no-rain detection capability, response to extreme weather, analysis of error sources, and the dependence of product performance on the gauge density and altitude. Key findings reveal: (1) Both continuous and categorical metrics for IMERG V06 and V07 exhibit significant regional variability, with V07 outperforming V06 across all aspects. (2) While V07 and V06 demonstrate comparable skill in reproducing precipitation frequency during extreme events, V07 shows significantly higher consistency with gauge observations for duration-based indices, specifically maximum consecutive dry days (CDD (max)) and maximum consecutive wet days (CWD (max)). (3) Except for Late runs, error components of other datasets for IMERG V06 and V07 are consistently higher than the total error, with the false and missed bias being the dominant contributors. The systematic error of all datasets is significantly lower than the random error. (4) IMERG V07 outperforms V06 across most metrics in both flood and non-flood seasons. All products show enhanced skill in detecting precipitation events during the flood season, indicating a consistent seasonal dependency in performance. (5) Most evaluation metrics for both IMERG versions correlate positively with gauge density and negatively with altitude changes. However, V07 demonstrates a reduced sensitivity to these environmental factors compared to V06. In summary, IMERG V07 demonstrates significant performance enhancements over V06, yet further improvements are essential. These insights provide critical guidance for both users and data developers, promoting further advancements in precipitation estimation.