These are (going to be) benchmarks that look at the performance of features, that have been uniquely implemented in the library.
The benchmarks are run with criterium.
nemesis comes with its own parser, which, in some circles, may or may not be frowned upon. Nevertheless, the parser is consistently benchmarked against a number of other ones stemming from other popular JVM libraries:
Loading
Given nemesis esoteric approach to working with JSON, some may assume that its operations may be somewhat inefficient.
As such, each one of them is consistently benchmarked:
Loading