Oh, my, this is a fun one.
When writing for my (English-speaking) readers, I use "race" for both "subspecies" and "a sophont quote-species-unquote". Those who read my linguistic appendices and/or footnotes will find the note that those uses are translations of completely different words in the in-world language; English broken, please fix.
So why not just "species"? Well, in least to most important order:
1. Some species have big ol' sticks up their butts about being classified in the same way as non-sophont lifeforms. (Plus, it makes my wiki look untidy.)
2. Not all "race-as-sophont species" are actually species. In my 'verse, the ren-gu and ren-qu are a symbiont pair who are quite insistent upon their oneness despite their technical twoness...
...There are polytaxic "species" who are one culture/people/polity likewise, and are offended at being split up. There are the people of machine ascent who don't _have_ a concept of species, they have models and codelines.
3. And so on and so forth.
4. In short, it's easier to coin a word than offend a lot of people, most of whom have guns and some of whom are guns.
(And why not "people"? Well, while some of those groups are unquestionably the same "race-meaning-sophont-species" in their own eyes, they are definitely separated into different "peoples" in ways not dependent on phenotype. Just ask them, and try not to pick a side.)
@martinsuntals
(Those paying extra attention will also note the sub-footnote that the former isn't cognate to the way 'race' is used in English, either: biological races have to breed true, which human races don't. Humans have ethnies and clines, and the Imperial Exobiological Institute is more than pedantic enough to make this point repeatedly, and with emphatic hand gestures.)