We modeled several thermal evolution scenarios for Ceres to explore the nature of large, wet protoplanets and to predict current-day evidence that might be found by close inspection, such as by the Dawn mission. The density for Ceres is near 2.1, suggesting a water content between 17% and 27% by mass. Short- and long-lived radioactive nuclide heating is considered. Even if only long-lived radionuclide heating is assumed, the water ice in Ceres melts quickly and a water mantle forms, but an approximately 10-km crust does not melt. The circulating warm water would alter the silicates. As heat is lost by conduction through the frozen crust, water begins to freeze out at the base of the crust. Solid-state convection begins and transports more heat as well as perhaps material dissolved or entrained in the water to or near the surface. Ceres' water layer eventually (but perhaps not entirely) freezes, forming a layered density structure with perhaps some liquid water remaining today. Our differentiated models are in agreement with the recently measured difference between the equatorial and polar radii. We find that Ceres' existence and evolution depend critically on it containing water at formation, and this depends strongly on the combination of when it accreted and the amount of 26Al present in the pre-Ceres ~1-km-sized objects; slightly more 26Al or earlier accretion produces a dry Vesta-like object. Melting and freezing plus mineralization would lead to several dimensional changes over time, creating topographic features, zones of weakness, and perhaps disruptions in the crust.