FV3 was configured as a global model with a stretching grid providing higher resolutions over CONUS; the mean grid spacing over the globe is about 13 km. A two-way interactive grid of about 3 km grid spacing was nested inside the FV3 global grid to further enhance the resolution over the CONUS region. Forecasts were run every day of the 2017 HWT EFP (Monday through Friday, 1 May through 2 June; 25 cases) and were initialized from the operational T1534 GFS analysis at 00Z each day and run for 120 hours. The CAPS FV3 forecasts are evaluated using neighborhood-based and object-based forecast evaluation metrics for hourly accumulated precipitation and instantaneous reflectivity. They are also compared with those of the GFDL forecasts, and with the control member of the 3-km CAPS Storm Scale Ensemble Forecasts (SSEF) that employed the WRF ARW model with Thompson microphysics. In addition to precipitation forecasting, other aspects including the prediction of updraft helicity, dryline position, and boundary layer structures will also be examined. Preliminary evaluations indicate that the precipitation and reflectivity forecasts are generally comparable with forecasts of other convection-permitting models presented at HWT, although there are also aspects that require improvement. Detailed results will be reported at the conference.