Genesis 1:5 uses a cardinal number "one" to describe the first day - יֹ֥ום אֶחָֽד - whereas all of the following days are modified by ordinals (WLC):
1:8 - יֹ֥ום שֵׁנִֽי
1:13 - יֹ֥ום שְׁלִישִֽׁי
1:19 - יֹ֥ום רְבִיעִֽי
1:23 - יֹ֥ום חֲמִישִֽׁי
1:311 - יֹ֥ום הַשִּׁשִּֽׁי
2:2 - בַּיֹּ֣ום הַשְּׁבִיעִ֔
The LXX follows suit (ἡμέρα μία...ἡμέρα δευτέρα...).
I would expect to see רִאשֹׁ֔ון instead of אֶחָֽד in v. 5.
HALOT does mention "first" as a possible meaning of אֶחָֽד2 but only for dates3 plus Gen 1:5 and 1 Sam 1:2.4 The latter is translated "one" in many (most) translations. Most translate אֶחָֽד in Genesis 1:5 as "first" (ESV, NIV, NET, KJV...), but a few do say "one" (NASB, RSV). The latter reads in English to me as if the previous phrase:
...and there was evening, and there was morning...
is defining "one day." On the other hand, the translation "the first day" seems to name the contents of the past "few"5 verses.6 Maybe that isn't so much different, but the contrast with the rest of the days numbered ordinally seems odd. Is there any significance to this?
1. Why on day 6 it's suddenly time for the article, and then on day 7 time for two of them (reverting to one when restated in the following verse), remains a mystery to me. Perhaps a topic for another question.
2. BDAG includes a similar usage of εἷς, "perhaps Hebraistic".
3. This has also been addressed in brief on BH.SE here where one answer indicates that it's always cardinal but the accepted answer refers to the ordinal usage with dates as in HALOT.
4. This paper also points out another use of אחד as an ordinal with "countables" (probably also in HALOT but I'm not too good with it), but, in distinction to Gen 1:5, every time with the article. (This paper also provides a partial answer to my question, but I'm not sure it's the whole story since he ends up disagreeing with most modern translations.)
5. I'm not committing, but this is addressed here.
6. In English we also have the option "day one," which strikes this native speaker as weirdly closer to "first day" than it is to "one day". Apparently we also don't have strict differentiation between cardinals and ordinals. Perhaps similar to the countables noted in #4 above.