lev¶

Introduction¶

lev is an extremely fast Python library for the Levenshtein distance and similarity ratio, written in Rust. 🦀

Installation¶

Install the latest version from pypi:

uv add lev-rs

or if you prefer slow:

pip install lev-rs

Usage¶

import lev

lev.distance("kitten", "sitting")   # 3
lev.distance("résumé", "resume")    # 2
lev.distance("日本語", "日本")       # 1

lev.ratio("kitten", "sitting")      # 0.769...
lev.ratio("", "")                   # 1.0

For more details on the API see the API Reference.

Benchmarks¶

lev is benchmarked against the fastest Python Levenshtein libraries: rapidfuzz, editdistance, and edlib. We excluded slower implementations like pylev and python-Levenshtein.

Benchmarks

Benchmarks were run on an Apple Mac Mini M2 Pro (macOS 26.2) using Python 3.13. The ASCII, Latin-1, CJK, and Emoji pairs are exactly 100 characters long; the realistic-text pair is natural-length prose. Results represent the total wall time for 1,000 repetitions using Python's timeit. To reproduce, run uv run scripts/benchmark.py.

ASCII¶

lev is significantly faster than the other libraries on 100-character ASCII strings.

ASCII benchmark – light ASCII benchmark – dark

Other Encodings¶

lev maintains its lead across all four CPython string-encoding kinds.

Latin-1¶

Latin-1 benchmark – light Latin-1 benchmark – dark

CJK¶

CJK benchmark – light CJK benchmark – dark

Emoji¶

Emoji benchmark – light Emoji benchmark – dark

Realistic Text¶

This pair is a short customer support message with a handful of natural typos.

Realistic text benchmark – light Realistic text benchmark – dark