In my personal opinion, it's worth a while to just drop the support without asking. Python 3.7 and 3.8 users would just use last compatible version.

I don't know if you have any statistics on pypi.org for which python version package has been requested though. If you have such statistics, it's worth a while to look into that

Thanks @eirnym for your ideas! I used this to get a clearer picture of the downloads, so I've used pypistats. This is the result:

$ pypistats python_minor semver
┌──────────┬─────────┬─────────────┐
│ category │ percent │   downloads │
├──────────┼─────────┼─────────────┤
│ 3.9      │  26.82% │  39,537,839 │
│ 3.10     │  17.87% │  26,341,633 │
│ 3.11     │  16.78% │  24,733,356 │
│ 3.6      │  13.58% │  20,010,727 │
│ 3.12     │   7.62% │  11,235,062 │
│ null     │   7.19% │  10,602,594 │
│ 3.8      │   7.01% │  10,335,207 │
│ 3.7      │   2.40% │   3,543,235 │
│ 3.13     │   0.56% │     822,993 │
│ 2.7      │   0.15% │     228,150 │
│ 3.14     │   0.01% │       8,658 │
│ 3.5      │   0.00% │       1,293 │
│ 3.4      │   0.00% │           4 │
│ Total    │         │ 147,400,751 │
└──────────┴─────────┴─────────────┘

Date range: 2024-07-25 - 2025-01-21

As "category" refers to the Python version, it looks like 3.7 and 3.8 is still being used. Maybe I consider of removing 3.7.

Nice statistics.

Python 3.8 will gone soon as well quite soon, it's just after EOL and not every venv is switched over to higher python versions. (As you can observe even in Python 2.7 usage)

Thus, in my personal opinion, I'd keep support for minor version, but drop for major version. This would ease your development process quite a lot.