Putting aside the R vs Python question (as as noted in this thread, you can use R in a Jupyter notebook and Python in an RMarkdown notebook), I much prefer RMarkdown notebooks. RMarkdown notebooks are plain text, so you can read them easily in any text editor (which also means they play well with git, unlike Jupyter notebooks).
And it's meant to work with the RStudio IDE, so I get a much more seamless experience going between regular code and notebooks (although this is admittedly a more R-centric benefit, at least until and unless RStudio adds Python support outside of notebooks).
A benefit I like about rmarkdown is that it makes it very easy for me to create templated reports. They're built in a way that makes it easy for me to work either iteratively (due to caching of blocks) or rerun the whole thing and get an output.
And it's meant to work with the RStudio IDE, so I get a much more seamless experience going between regular code and notebooks (although this is admittedly a more R-centric benefit, at least until and unless RStudio adds Python support outside of notebooks).