Apps · Python library

Sentence Transformers

by Hugging Face · Apache-2.0 · 19,219 GitHub stars

The standard Python library for cross-encoder rerankers; with Transformers pipelines it runs every classifier and reward model here.

macOSWindowsLinux

Get Sentence Transformers

At a glance

Best for
Scripts and notebooks: a few lines of Python to score, rerank or classify on your own machine.
Inference engine
  • PyTorch (Transformers)
GPU support
  • NVIDIA CUDA
  • AMD ROCm
  • Apple Metal
  • CPU only
Opens
  • Safetensors (original weights)
  • ONNX
Features
  • HTTP API
Platforms
Python 3 on macOS (Apple MPS), Windows and Linux (CUDA, ROCm or CPU).
Speed
Eager PyTorch: simple, but slower per call than TEI or vLLM; batch many pairs at once.

Sources: github.com

Install

macOS

pip install sentence-transformers

Windows

pip install sentence-transformers

Linux

pip install sentence-transformers

57 models on this site run in Sentence Transformers

Decision models it can load: rerankers, classifiers and scorers with a supported architecture. Open a model for its speed and memory on your hardware.

Model Made by Type Size Opens it as Quality tier
Qwen3 Reranker 4B Alibaba Decision 4.0B Original weights 90
Qwen3 Reranker 8B Alibaba Decision 8.2B Original weights 90
Skywork Reward V2 Llama 3.1 8B Skywork Decision 7.5B Original weights 90
zerank-2 ZeroEntropy Decision 4.0B Original weights 88
CompassJudger-2 32B OpenCompass Decision 33B Original weights 86
Jina Reranker v3.5 Jina AI Decision 600M Original weights 86
Qwen3Guard Gen 8B Alibaba Decision 8.2B Original weights 86
Shieldstral 1.0 3B Mistral AI Decision 3.9B Original weights 86
Granite Guardian 4.1 8B IBM Decision 8.4B Original weights 84
Jina Reranker v3 Jina AI Decision 600M Original weights 84
Llama 3.1 Nemotron Safety Guard 8B v3 NVIDIA Decision 8.0B Original weights 84
Qwen3Guard Gen 4B Alibaba Decision 4.4B Original weights 84
Skywork Reward V2 Qwen3 8B Skywork Decision 7.6B Original weights 84
mxbai-rerank-large v2 Mixedbread Decision 1.5B Original weights 82
gpt-oss-safeguard 120B OpenAI Decision 117B (5.1B active) Original weights 80
Granite Guardian 3.3 8B IBM Decision 8.2B Original weights 80
Qwen3 Reranker 0.6B Alibaba Decision 600M Original weights 80
Selene 1 Mini 8B Atla Decision 8.0B Original weights 80
WildGuard 7B Ai2 Decision 7.3B Original weights 80
zerank-1 small ZeroEntropy Decision 1.7B Original weights 80
DynaGuard 8B University of Maryland Decision 8.2B Original weights 79
BGE Reranker v2 Gemma BAAI Decision 2.5B Original weights 78
CompassJudger-2 7B OpenCompass Decision 7.6B Original weights 78
Nemotron Content Safety Reasoning 4B NVIDIA Decision 4.3B Original weights 78
OpenAI Privacy Filter OpenAI Decision 1.4B (50M active) Original weights 78
gpt-oss-safeguard 20B OpenAI Decision 22B (3.6B active) Original weights 76
Llama Prompt Guard 2 86M Meta Decision 280M Original weights 76
mxbai-rerank-base v2 Mixedbread Decision 490M Original weights 76
NuExtract 2.0 4B NuMind Decision 3.8B Original weights 76
Prometheus 2 8x7B Prometheus Eval Decision 47B (13B active) Original weights 74
Qwen3Guard Gen 0.6B Alibaba Decision 750M Original weights 74
BGE Reranker v2 M3 BAAI Decision 570M Original weights 72
DeBERTa v3 Large Zero-shot v2.0 Moritz Laurer Decision 440M Original weights 72
DuoGuard 0.5B DuoGuard Decision 490M Original weights 72
Skywork Reward V2 Qwen3 0.6B Skywork Decision 600M Original weights 72
DynaGuard 1.7B University of Maryland Decision 2.0B Original weights 70
GTE Reranker ModernBERT Base Alibaba Decision 150M Original weights 70
Jina Reranker v2 Base Multilingual Jina AI Decision 280M Original weights 70
ModernBERT Large Zero-shot v2.0 Moritz Laurer Decision 400M Original weights 70
NuExtract 2.0 2B NuMind Decision 2.2B Original weights 70
Granite Embedding Reranker English R2 IBM Decision 150M Original weights 68
Llama Guard 3 8B Meta Decision 8.0B Original weights 68
Llama Prompt Guard 2 22M Meta Decision 71M Original weights 68
Prometheus 2 7B Prometheus Eval Decision 7.2B Original weights 68
GLIDER Patronus AI Decision 3.8B Original weights 66
LettuceDetect Large KR Labs Decision 400M Original weights 66
Llama Guard 4 12B Meta Decision 12B Original weights 66
ProtectAI DeBERTa v3 Prompt Injection v2 Protect AI Decision 180M Original weights 66
ShieldGemma 27B Google Decision 27B Original weights 66
DeBERTa v3 Base Zero-shot v2.0 Moritz Laurer Decision 180M Original weights 64
Flow Judge v0.1 Flow AI Decision 3.8B Original weights 64
ShieldGemma 9B Google Decision 9.2B Original weights 64
HHEM 2.1 Open Vectara Decision 110M Original weights 62
ShieldGemma 2B Google Decision 2.6B Original weights 58
BART Large MNLI Meta Decision 410M Original weights 56
Llama Guard 3 1B Meta Decision 1.5B Original weights 56
ms-marco MiniLM L6 v2 Sentence Transformers Decision 23M Original weights 52

Alternatives

LM Studio Ollama llama.cpp vLLM Jan Text Embeddings Inference MLX LM SGLang GLiNER Lemonade Infinity Unsloth Desktop Docker Model Runner LocalAI Atomic Chat

Which app should I use?

Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.