If you're working with json objects with sizes on the higher end quite often you're not going to need the entirety of them, just a small part of them. If that is the workload what then to do is simply parse as little data as possible: skip the validation, locate the relevant bits, and then start parsing, validation and all the stuff. In this optimizing the json scanner/lexer gives much greater improvement than optimizing the parser.

Though this job is trickier than it may look. The logic to extract the "relevant" bits is often dynamic or tied to user input but for the scanner/lexer to be ultrafast it has to be tightly compiled. You can try jitting but libllvm is probably too heavyweight for parsing json.