I would suggest camera module 2 or 3 not HQ. Mostly as you don't need 2028x1520 but slower recording speed and/or streaming video speed. 1080p is good enough. Having 60 fps is most likely better than 40 but that is just me. How high they both go is also relative to how high you sent the resolution. module 3 can do 120 when 640x480 and HQ can do 120 when at 1332x920.

Unless you already got it I would go with module 2 for your use case. As it isn't that much worse than 3 and more compatible.

Here is a list of comparisons.

Camera module 3 vs Camera module 2

The camera module 3
pros:
1080p 60hz record though software might limit you to slower
Better color reproduction

cons:
weights more by 1 gram being 3 grams of weight ( unsure if 1 gram is a lot of a little for your set up)
requires more space on the vertical axis
ONLY compatible with the modern libcamera software environment and not the legacy closed-source camera stack

The camera module 2
pros:
supports 1080p 47frames per second though quickest at 240p at 207frames per second
generates smaller files
color reproduction acceptable for your use case
cheaper
lighter at 3 grams instead 4 grams
is compatible with both the legacy closed-source camera stack and modern libcamera software environment
physically thinner than camera module 3

cons:
It does not support 720p
End of life is sooner though still well supported and as it is open source end of life will just apply to hardware not software
have to manually adjust the camera when moving it physically

Camera module 3 vs Camera module HQ

Camera module HQ
pros:
2028x1520 at 40 frames per second

con:
physically bigger at 38mm x 38mm x 18.4mm not including the lens
you need a external lends for it to be better than the camera module 3
more expensive at 50 USD before you but the lens

unsure about compatibility with software

​

Camera module 3
pros:
faster record at 1080p at 60fps
better color reproduction
weights less how much depends on which lens you use with the module HQ
cheaper
physically smaller at 25mm x 24mm x12.4mm

con:
doesn't support as many lens
only goes up to 1080p not higher