The top 3 sides are pressure inlets/oulets with 0 gauge pressure specified, and the bottom 3 are all walls. Like a simplified model for a cup of water surrounded by dry air. I've added images of velocity magnitude, temperature, mesh, and steam concentration after 7 s of flowtime below.

The flow and steam concentration after a short time look reasonable. The temperatures maybe aren't diverging so fast, but they are lower in the vapor region where they shouldn't be. The liquid water should be cooling due to evaporation, not the steam.

I also found that it's probably related to enthalpy transfer from phase change. Modifying the standard state enthalpy of the water materials changes the temperature drops accordingly (delta H higher = more T drop in vapor, and vice versa). While I expect to see the effects of evaporative cooling, I feel the trend is opposite what it's supposed to be - steam should not be cooling due to evaporation, the liquid water should be. I'm not sure why it would be reversed like this.

Thanks again for your help.