The Canadian Surface and Precipitation Reanalysis (CaSR; https://hpfx.collab.science.gc.ca/~scar700/rcas-casr), developed by Environment and Climate Change Canada (ECCC), is a high-resolution (10 km) reanalysis spanning more than 56 years (1968–2024) over North America. CaSR is produced using a sequential, coupled modeling framework that integrates ECCC’s operational surface and precipitation data assimilation systems along with numerical weather prediction systems and extensive in situ surface observation networks. Version 3.2 represents a major upgrade over the previous release (v2.1). Key improvements include the transition from ERA-Interim to ERA5 for initial and boundary conditions, an upgrade of the GEM atmospheric model from version 4.8 to 5.2, and refinements to the ISBA land surface scheme, particularly in the coupling between snowpack processes and the surface energy balance. In addition, the assimilation of in situ surface observations has been enhanced. Together, these changes lead to substantial improvements, with error reductions for temperature- and wind-related variables ranging from a few percent to approximately 25% across seasons and Bukovsky subdomains, and a marked reduction of the persistent cold bias during the spring snowmelt period. This presentation describes the CaSR methodology and highlights the main features and performance of version 3.2 through comparisons with observational datasets, the previous CaSR version, and other major reanalyses. It also illustrates the impact of these improvements on selected downstream applications such as hydrology. The presentation concludes with an overview of ongoing developments, and the roadmap toward the next CaSR version, planned for release in 2028.