Volcanic stratospheric aerosols represent a key forcing for climate simulations in Phase 7 of the Coupled Model Intercomparison Project (CMIP7). The CMIP7 stratospheric aerosol forcing datasets consist of: i) a volcanic sulfur emission inventory; and ii) stratospheric aerosol optical properties, derived from emissions using a reduced-complexity volcanic aerosol model for 1750–1978, and from satellite observations for 1979–2023. Here, we compare the CMIP7 and CMIP6 stratospheric aerosol forcing datasets over the historical period, and evaluate them against an extensive compilation of observational datasets, including lunar eclipse, stellar extinction, pyrheliometers, lidar, and satellite datasets. CMIP6 and CMIP7 show reasonable agreement (< 20 % difference) on peak stratospheric aerosol optical depth (SAOD) for most large-magnitude eruptions including Krakatau (1883), Katmai (1912), Agung (1963), El Chichón (1982), and Pinatubo (1991). For small-to-moderate eruptions, SAOD is likely underestimated in CMIP6 during periods when no such eruptions are included, such as 1940–1960, while CMIP7 includes more pre-satellite small-to-moderate eruptions, as supported by lunar eclipse data. Observational datasets indicate that SAOD perturbations for some moderate-magnitude eruptions, including the 1902, 1921, 1928, 1929, and 1974 eruptions, are captured in CMIP6, but likely underestimated in CMIP7. In addition, independent lidar and stellar extinction measurements show that both CMIP6 and CMIP7 underestimate SAOD over the Northern Hemisphere after the 1982 El Chichón eruption due to observational gaps in satellite instruments. Our findings highlight the potential value of combining CMIP6's use of available observational constraints with CMIP7's extensive emission inventory to refine stratospheric aerosol forcing in future CMIP phases.