> if we left that up to popular opinion
... then national polls are not relevant, because marriage laws are set at the state level. In fact, until Obergefell v. Hodges, it essentially was left up to popular opinion, and many people lived in states where it was unpopular enough to pass legislative, or even constitutional bans.
But no matter how you slice it, I see no obvious support in the data for the original claim that "gay marriage was legalized when popular support was only at 25%".
Now maybe there were some states that had support at only 25% in 2015. For example, https://www.nbcnews.com/news/us-news/where-same-sex-marriage... is a 2014 article that mentions that Missisipi had 14% support in 2004 and support went up by "1-2% per year since then". So _maybe_ it was at only 25%, as an absolute lower bound. But chances are it was higher than that even there.
Support for gay marriage was at the 25% level nationally in ... well, per https://web.archive.org/web/20050517033538/http://www.aei.or... page 21 I would put it in the mid-1990s, 20 years before full nationwide legalization. Of course legalization in various states preceded national legalization, as expected.
The other things you list predate modern polling (except maybe Black voting rights, depending on how you define that). But just looking at women's suffrage, it was a constitutional amendment. That means is got 2/3 majorities in both houses of congress and ratification by 3/4 of state legislatures. It's hard to see how something with only 25% popular support could manage that, so at this point I'm going to ask you for data to back up that extraordinary claim.