I have just published a project on GitHub with an experiment to compare the speed from different Scala libraries to de-serialize JSON data into case classes. I have been using JSON reading and writing not only for web, but also data analysis in general, and I was afraid some of the libraries might still have performance issues, what apparently was usual in the past couple of years.
The experiment has showed that all libraries are pretty mature today, with no obvious winner or loser in the experiment. The most interesting is that apparently the performance might be very dependent on the specific model you are reading, what indicates there may be really little room for speed improvements left.
Move to my GitHub project to look at the code and see the results!