Rust Parsing Benchmarks

This repo tries to assess Rust parsing performance.

crate	parser type	action code	integration	input type	precedence	parameterized rules	streaming input
chumsky	combinators	in source	library	&str, &[u8], custom	pratt	Yes	Yes
combine	combinators	in source	library	&str	?	?	?
grmtools	CFG	in grammar	library	?	?	?	?
lalrpop	LR(1)	in grammar	build script	&str	none	Yes	No
lelwel	LL(1)	in grammar	build script	&str	pratt	Yes	No
logos	lexer	in source	proc macro	&str, &[u8]	?	?	?
nom	combinators	in source	library	&str, &[u8], custom	pratt	Yes	Yes
parol	LL(k)/LALR(1)	in source	build script	&str	climbing	No	No
peg	PEG	in grammar	proc macro (block)	&str, &[T], custom	climbing	Yes	No
pest	PEG	external	proc macro (file)	&str	climbing	No	No
winnow	combinators	in source	library	&str, &[T], custom	none	Yes	Yes
yap	combinators	in source	library	&str, &[T], custom	none	Yes	?

Formerly, we compared:

• pom: lack of notoriety

Results

Name	Overhead (release)	Build (debug)	Parse (release)	Downloads	Version
null	0 KiB	249ms	3ms	-	-
grmtools	2,610 KiB	11s	164ms		v0.13.10
chumsky	150 KiB	4s	32ms		v0.10.1
combine	181 KiB	4s	53ms		v3.8.1
lalrpop	1,523 KiB	11s	37ms		v0.22.2
lelwel	142 KiB	4s	10ms		v0.8.0
logos	90 KiB	4s	21ms		v0.15.0
nom	98 KiB	3s	65ms		v8.0.0
parol	492 KiB	9s	174ms		v4.1.0
peg	80 KiB	2s	23ms		v0.8.5
pest	130 KiB	5s	61ms		v2.8.1
serde_json	59 KiB	3s	14ms		v1.0.142
winnow	75 KiB	2s	28ms		v0.7.12
yap	61 KiB	527ms	33ms		v0.12.0

System: Linux 6.8.0-62-generic (x86_64), rustc 1.89.0 (29483883e 2025-08-04) w/ -j 8

Note:

• For more "Parse (release)" comparisons, see parser_benchmarks
• Parsers have not been validated and might have differing levels of quality (#5)

Running the Benchmarks

$ ./bench.py
$ ./format.py