simsimd
Portable mixed-precision BLAS-like vector math library for x86 and ARM
Decision gist · record as of 2026-08-14
Yes. SimSIMD is actively maintained, has no dependencies, and offers significant performance improvements over standard libraries like scipy and NumPy for vector operations. The permissive Apache-2.0 license and broad platform support (x86, ARM, multiple Python versions) make it a low-risk addition to performance-critical workloads. Install it if vector similarity or distance computation is a measurable bottleneck in your application.AI-flagged interpretation of the facts on this page — verify before relying
Before you install
- Medium install friction due to compiled wheels for multiple platforms.
- The package is actively maintained with a recent release and no runtime dependencies, making it straightforward to integrate once installed.
License · maintenance · safety
Apache-2.0 (permissive) — Licensed under Apache-2.0 (permissive), allowing use in commercial and proprietary projects without significant legal constraints.
last release 2026-03-07 (160 days) · last repo commit 2026-08-07 · 1,874 stars
0 known vulnerabilities (OSV.dev, 2026-08-14) · 3,762,725 downloads/mo, #2,501 on PyPI
Alternatives
Verify before relying
pip install simsimd
import simsimd
# Compute cosine distance between two float32 vectors
dist = simsimd.cosine(a, b)- Whether Python 3.9 is actually supported despite appearing in classifiers but not in friction_evidence wheels
- Performance gains relative to scipy.spatial.distance on typical workloads outside the benchmarked 1536d case
- Whether complex vector support is available through the Python API or only in C
What it is and what it does
SimSIMD is a mixed-precision vector math library that implements over 350 SIMD-optimized kernels for computing distances and similarities between vectors. It supports float64, float32, float16, bfloat16, int8, int4, and binary vectors, as well as sparse and complex vectors. The library is designed to replace or supplement BLAS level-1 operations and scipy.spatial functions with significantly faster implementations that leverage modern CPU instruction sets (AVX-512, SVE, NEON) on both x86 and ARM architectures.
The package ships as pre-compiled wheels for multiple platforms and Python versions, with no runtime dependencies. It handles operations like Euclidean and cosine distances for vector search, dot-products for DSP and quantum computing, Hamming and Jaccard distances for bit-level comparisons, set intersections for sparse vectors, Mahalanobis distance for scientific computing, and divergence measures for probability distributions. The library is commonly used in AI, search, and database workloads where vector similarity computation is a performance bottleneck.
Use it for
- Vector search and similarity retrieval in semantic search or embedding-based recommendation systems
- Batch computation of pairwise distances in clustering or nearest-neighbor algorithms
- Mixed-precision inference in AI pipelines where bfloat16 or float16 vectors need fast similarity scoring
- Sparse vector operations and set intersections for text analysis or information retrieval
- Geospatial distance calculations and scientific computing with Mahalanobis distance or quadratic forms
Worth the install?
AI-flagged interpretation of the facts on this page. Verify before relying on it.
Yes.
SimSIMD is actively maintained, has no dependencies, and offers significant performance improvements over standard libraries like scipy and NumPy for vector operations. The permissive Apache-2.0 license and broad platform support (x86, ARM, multiple Python versions) make it a low-risk addition to performance-critical workloads. Install it if vector similarity or distance computation is a measurable bottleneck in your application.
Install
simsimd on PyPI
Before you install
Medium install friction due to compiled wheels for multiple platforms. The package is actively maintained with a recent release and no runtime dependencies, making it straightforward to integrate once installed.
License in practice
Licensed under Apache-2.0 (permissive), allowing use in commercial and proprietary projects without significant legal constraints.
Quickstart
pip install simsimd
import simsimd
# Compute cosine distance between two float32 vectors
dist = simsimd.cosine(a, b)Verify before relying
- Whether Python 3.9 is actually supported despite appearing in classifiers but not in friction_evidence wheels
- Performance gains relative to scipy.spatial.distance on typical workloads outside the benchmarked 1536d case
- Whether complex vector support is available through the Python API or only in C
Package facts
| License | Apache-2.0 permissive |
| Python support | Not specified |
| Install friction | Medium. Platform-specific wheel |
| Runtime dependencies | None |
| Maintenance | Actively maintained 160 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 3,762,725 / month, #2,501 on PyPI 30-day window, as of 2026-08-14 |
| Known vulnerabilities | None known OSV.dev, checked 2026-08-14 |
| Classifiers | Development Status :: 5 - Production/StableIntended Audience :: DevelopersIntended Audience :: Information TechnologyLicense :: OSI Approved :: Apache Software LicenseNatural Language :: EnglishOperating System :: MacOSOperating System :: Microsoft :: WindowsOperating System :: POSIX :: LinuxProgramming Language :: CProgramming Language :: Python :: 3Programming Language :: Python :: 3.10Programming Language :: Python :: 3.11Programming Language :: Python :: 3.12Programming Language :: Python :: 3.13Programming Language :: Python :: 3.14Programming Language :: Python :: 3.9Programming Language :: Python :: Free Threading :: 3 - StableProgramming Language :: Python :: Implementation :: CPythonTopic :: Scientific/Engineering :: Artificial IntelligenceTopic :: Scientific/Engineering :: Bio-InformaticsTopic :: Scientific/Engineering :: ChemistryTopic :: Scientific/Engineering :: Information AnalysisTopic :: Scientific/Engineering :: Mathematics |
Evidence: simsimd-6.5.16-cp310-cp310-macosx_10_9_x86_64.whl; simsimd-6.5.16-cp310-cp310-macosx_11_0_arm64.whl; simsimd-6.5.16-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; simsimd-6.5.16-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl; simsimd-6.5.16-cp310-cp310-musllinux_1_2_aarch64.whl; simsimd-6.5.16-cp310-cp310-musllinux_1_2_x86_64.whl; simsimd-6.5.16-cp310-cp310-win_amd64.whl; simsimd-6.5.16-cp310-cp310-win_arm64.whl; simsimd-6.5.16-cp311-cp311-macosx_10_9_x86_64.whl; simsimd-6.5.16-cp311-cp311-macosx_11_0_arm64.whl; simsimd-6.5.16-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; simsimd-6.5.16-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl; simsimd-6.5.16-cp311-cp311-musllinux_1_2_aarch64.whl; simsimd-6.5.16-cp311-cp311-musllinux_1_2_x86_64.whl; simsimd-6.5.16-cp311-cp311-win_amd64.whl; simsimd-6.5.16-cp311-cp311-win_arm64.whl; simsimd-6.5.16-cp312-cp312-macosx_10_13_x86_64.whl; simsimd-6.5.16-cp312-cp312-macosx_11_0_arm64.whl; simsimd-6.5.16-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; simsimd-6.5.16-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl
Tags
Let your AI agent find packages like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 14,416 PyPI packages by what they can do, searchable in plain language.
wish › “vector similarity and distance computation”
- simsimdSimSIMD provides SIMD-optimized kernels for computing vector…
- python-LevenshteinComputes Levenshtein edit distance, string similarity, and…
- pgvectorAdds vector storage and similarity search support to PostgreSQL…
Give your agent the search over MCP, or paste the wish link into any chat.
More Artificial Intelligence packages
LiteLLM provides a unified Python interface to call 100+ LLM providers (OpenAI, Anthropic, Gemini, Bedrock, Azure, and others) using OpenAI-compatible API format, available as both a Python SDK and a self-hosted AI Gateway proxy server.
Install it if you need to work with multiple LLM providers or want to centralize LLM routing in your organization.
Client library and CLI tool for downloading, uploading, and managing models, datasets, and repositories on the Hugging Face Hub platform.
Install it if you work with Hugging Face Hub models or datasets.
LangChain provides a framework for building agents and LLM-powered applications by composing language models, tools, and memory through a unified API that abstracts over multiple model providers.
hf-xet provides chunk-based deduplication and efficient file transfer for the Hugging Face Hub, enabling faster uploads and downloads of large files with local disk caching.
Tokenizers converts raw text into token sequences for NLP models, with support for training custom vocabularies and using pre-built tokenizers (BPE, WordPiece) optimized for speed via Rust.
Transformers provides a unified framework for loading, fine-tuning, and running state-of-the-art pretrained models across text, vision, audio, video, and multimodal tasks using PyTorch, JAX, or TensorFlow.
Install it if you need to run or train any transformer-based model for NLP, vision, audio, or multimodal tasks.
See also numkong · usearch · faiss-gpu · scann · scikit-misc · annoy · mkl · pysimdjson · stringzilla · traittypes