That depends. If you are an expert in the area you can pretty much find out the important information in a well-written abstract and conclusion.
For most papers I find that it takes a single read, partially to pick up new ideas, techniques, and overviews of related research.
For some papers I have tried several times, especially the poorly written ones that contain a lot of noise, particularly bad in the logic area with pages of unexplained judgments.
And there are piles of papers I simply don't have the time and background to understand.
I find the hardest and most time consuming part is reading the references. I try to find and read some prior papers from the bibliography. For really interesting research this turns into a book-sized pile of reading. With enough experience in the area I usually have already read the most interesting references.
The other hard part is the volume of published papers. It feels like sipping from a firehose.
