It depends on the property you want to observe or study. For chiral compounds, DFT is often used to predict the vibrational and circular electronic dichroism. You can model both enantiomers separately and then together and observe the differences in the spectra.

Enantiomers have the same energy (OK, there is a parity-violating energy difference but that is so small as to be irrelevant to us chemists). In the absence of a chiral environment, there is no difference in stability between enantiomers.

If you do decide to separately calculate the energies of two enantiomers, compare them, and obtain a difference in stability, this is telling you something about your computational methodology:

- are you looking at the same (mirror-image) conformers?

- is the DFT numerical integration grid too coarse?

- how about optimization and SCF convergence criteria?

If you are studying a racemic mixture, it is generally a good idea to carry out DFT calculations on both enantiomers separately. This will allow you to compare the electronic and magnetic properties of the two enantiomers and determine if there are any significant differences between them.

If there is a significant difference between the theoretical results for the two enantiomers, it may be possible to hypothesize which stereoisomer is more stable or active. However, it is important to keep in mind that DFT calculations are a theoretical predictions and may not always agree with experimental results. Therefore, it is important to verify the predictions of DFT calculations with experimental measurements, such as spectroscopy or X-ray diffraction.

Additionally, It's worth noting that, In certain cases, the difference in stability or activity between the two enantiomers may be small and may not be easily distinguished by DFT calculations. In such cases, other theoretical or experimental methods such as ab-initio calculations, Quantum Monte Carlo, or kinetic studies, may be needed to determine which enantiomer is more stable or active.