# erikbern/ann-benchmarks

Benchmarks of approximate nearest neighbor libraries in Python

Repository: https://github.com/erikbern/ann-benchmarks
Canonical: https://ross.abutalabs.com/products/ann-benchmarks
Homepage: http://ann-benchmarks.com
Language: Python
License: MIT
License Family: permissive
Topics: nearest-neighbors, benchmark, docker
Last push: 2026-07-10T08:44:55+00:00

## Health v2 (maintenance only)
Score: 73/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 91, release rhythm 35, longevity 100
- inputs: {"age_days": 4115, "days_push": 54, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5717, forks 903 (observed 2026-08-28T04:09:28.224476+00:00)

## What it is
A benchmarking environment for approximate nearest neighbor (ANN) search algorithms, with pre-generated HDF5 datasets and Docker containers for each evaluated library. It publishes recall vs queries-per-second results across many datasets and distance metrics.

## Use cases
- compare approximate nearest neighbor libraries
- benchmark vector search algorithms
- evaluate faiss vs hnswlib vs annoy performance
- measure recall and queries per second for ANN indexes
- find the fastest nearest neighbor search for my dataset
- benchmark similarity search on standard datasets

## When to choose
- you need historical, reproducible ANN benchmark results across many algorithms
- you want a ready-made harness with datasets and Docker containers for evaluating a new ANN method
- you need recall/QPS plots across distance metrics like angular, euclidean, hamming, and jaccard

## When to avoid
- you need an actively maintained benchmark - the project is no longer maintained and points to alternatives like VIBE
- you need production vector search rather than benchmarking
- you need benchmarks of very recent algorithms not included in the evaluated set

## Facets
- artifact type: library
- maturity: abandoned
- function: benchmarking, search-engine, vector-database
- domain: machine-learning, databases, performance
- platform: python
- tags: ann, nearest-neighbor-search, vector-search, hdf5-datasets, recall-vs-qps, algorithms, docker, linux

## Member repositories
- erikbern/ann-benchmarks (main) score 73

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:28.224476+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:53:28.522175+00:00, confidence not recorded.
  - readme: https://github.com/erikbern/ann-benchmarks (fetched 2026-08-28T04:09:28.224476+00:00, sha 6b62a599c04c)
  - homepage: http://ann-benchmarks.com (fetched 2026-08-29T08:48:57.865189+00:00, sha 9752fe975ebf)
- Data as of 2026-08-30T08:39:29.467469+00:00.