orjson

orjson is a fast, correct JSON library for Python. It benchmarks as the fastest Python library for JSON and is more correct than the standard json library or other third-party libraries. It serializes dataclass, datetime, numpy, and UUID instances natively. Despite not being 100% compatible, it is often used as replacement of than JSON functions in core Python 3. As of Jan 29th 2026 (https://github.com/ijl/orjson/commit/d58168733189f82b3fd0c05...) open issue tracker or pull requests has been removed from the orjson repo due to signal-to-noise ratio. Might be due to high amount of LLM-assisted noisy bug reports and/or PRs that are plagueing various projects on Github. It is unfortunate that even old issues won't be accessible for reference. What do you guys think?

I'm really surprised ijl got angry that his mail was quoted, it looks innocent enough to me. For reference it's been edited out here: https://github.com/pydantic/pydantic/issues/589 But github shows edits, so the edit is meaningless for privacy. Here's the original mail (yes, I'm blatantly ignoring his request to not publish this, I'm just this evil.) I've looked into replacing ujson in pydantic with orjson (https://github.com/ijl/orjson). In this implementation, the same JSON library is used for everything, and JSON outputs bytes without whitespace (as it's faster and JSON is a serialization format). If orjson is installed, it won't affect pydantic's benchmark for validation, but can be expected to improve whole-program performance. It's a large change with breaking changes to JSON methods, however, so ra

I asked about this on the Github issue regarding these benchmarks as well. I'm curious as to why libraries like ultrajson[0] and orjson[1] weren't explored. They aren't command line tools, but neither is pandas (which is benchmarked) right? Is it perhaps because the code required to implement the challenges is large enough that they are considered too inconvenient to use through the same way pandas was used (ie, `python -c "..."`)? [0] https://github.com/ultrajson/ultrajson [1] https://github.com/ijl/orjson

How big are those responses? I’ve seen terrible performance come down to serialization, which can be addressed by swapping in a fast serialization library like orjson (https://github.com/ijl/orjson). Though even then you’d probably have a hard time getting to tens of milliseconds. Other common culprits: poor indexing, n+1 queries.

And if you're looking for a fast JSON lib for CPython, orjson[1] (written in rust) is the best I've found. [1] https://github.com/ijl/orjson#performance

Not out of the box. For example, the fastest (and my favorite) JSON (de)serializer available for Python, orjson[1], does not support pypy. [1] https://github.com/ijl/orjson