paraphrase-MiniLM-L6-v2 by Sentence Transformers: Open Model

Runs On

What it takes to serve paraphrase-MiniLM-L6-v2 (23M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

Precision	Memory needed	Cheapest setup	Per hour	Also fits
16-bit	0.1 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
8-bit	0.0 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00
4-bit	0.0 GB	1x MI300X (192 GB) Vultr	$1.85	1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

SAVRN's Notes on paraphrase-MiniLM-L6-v2

Semantic search and clustering are the jobs here: the model turns a sentence or paragraph into a 384-dimensional vector, and everything downstream compares vectors. With 23 million parameters, the 16-bit memory need rounds to 0.1 GB and the weights round to zero. The cheapest configuration on file, one MI300X with 192 GB at $1.85 per hour on-demand, says the cost driver is not the model but how many documents you push through it.

Under Apache 2.0 you can embed proprietary documents, keep the vectors, fine-tune on your own sentence pairs and redistribute commercially, as long as the notices travel with it and significant changes are stated. Check the context length first: 512 tokens, so longer passages need chunking, and the chunking will shape results more than the model does. Then use a format the publisher already ships, safetensors, ONNX, OpenVINO, PyTorch or TensorFlow.

SAVRN Research, 2026-09-18

Model Card

By Sentence Transformers, published under apache-2.0, revision c9a2bfebc254.

This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.

Usage (Sentence-Transformers)

Using this model becomes easy when you have sentence-transformers installed:

pip install -U sentence-transformers

Then you can use the model like this:

from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]

model = SentenceTransformer('sentence-transformers/paraphrase-MiniLM-L6-v2')
embeddings = model.encode(sentences)
print(embeddings)

Usage (HuggingFace Transformers)

Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.

Read the full model card (316 words)

Configuration

Architecture: BertModel
Context length (tokens): 512
Layers: 6
Hidden size: 384
Feed-forward size: 1,536
Attention heads: 12
Vocabulary size: 30,522
Model type: bert

Identity and Version

Repository: sentence-transformers/paraphrase-MiniLM-L6-v2
Publisher: Sentence Transformers
Task: Sentence similarity
Modality: Text
Library: sentence-transformers
Parameters: 23M parameters
Languages: tf
Revision: c9a2bfebc254878aee8c3aca9e6844d5bbb102d1
First published: 2022-03-02
Last updated: 2025-03-06

Files and Weights

27 files, 886.0 MB in total. The weights are 14 files totalling 884.7 MB in bin, h5, onnx, safetensors.

Weights14 files · 884.7 MB

Configuration6 files · 1.3 KB

Tokenizer3 files · 697.9 KB

Documentation1 file · 3.5 KB

Other2 files · 579.6 KB

Repository1 file · 744 B

Every file

File	Type	Size	SHA-256
model.safetensors	Weights	90.9 MB	2ce4480dc3b2
onnx/model.onnx	Weights	90.4 MB	441a5dc61ff3
onnx/model_O1.onnx	Weights	90.4 MB	e4b50dcc09ca
onnx/model_O2.onnx	Weights	90.3 MB	d99151ccdfb7
onnx/model_O3.onnx	Weights	90.3 MB	053f24452815
onnx/model_O4.onnx	Weights	45.2 MB	891a31575319
onnx/model_qint8_arm64.onnx	Weights	23.0 MB	ccc4bb68331e
onnx/model_qint8_avx512.onnx	Weights	23.0 MB	ccc4bb68331e
onnx/model_qint8_avx512_vnni.onnx	Weights	23.0 MB	ccc4bb68331e
onnx/model_quint8_avx2.onnx	Weights	23.0 MB	7d2e9f418045
openvino/openvino_model.bin	Weights	90.3 MB	c6005063ac5c
openvino/openvino_model_qint8_quantized.bin	Weights	22.9 MB	f036c75118e1
pytorch_model.bin	Weights	90.9 MB	5d716de760ac
tf_model.h5	Weights	91.0 MB	ee09134d6f68
1_Pooling/config.json	Configuration	190 B	—
config.json	Configuration	629 B	—
config_sentence_transformers.json	Configuration	122 B	—
modules.json	Configuration	229 B	—
sentence_bert_config.json	Configuration	53 B	—
special_tokens_map.json	Configuration	112 B	—
README.md	Documentation	3.5 KB	—
openvino/openvino_model.xml	Other	211.3 KB	—
openvino/openvino_model_qint8_quantized.xml	Other	368.2 KB	—
.gitattributes	Repository	744 B	—
tokenizer.json	Tokenizer	466.1 KB	—
tokenizer_config.json	Tokenizer	314 B	—
vocab.txt	Tokenizer	231.5 KB	—

License and Download

License: apache-2.0
Access: Open weights, no gate
Download size: 884.7 MB

Download from Sentence Transformers

Released by Sentence Transformers through its official repository on Hugging Face. Read the license.

Built From

Described by arXiv:1908.10084

Memory Requirements

Precision	Weights in memory
As published	884.7 MB
16-bit	0.0 GB
8-bit	0.0 GB
4-bit	0.0 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Compare paraphrase-MiniLM-L6-v2

all-MiniLM-L6-v2 vs paraphrase-MiniLM-L6-v2

Questions About paraphrase-MiniLM-L6-v2

How much GPU memory does paraphrase-MiniLM-L6-v2 need?

About 0.1 GB at 16-bit and 0 GB at 4-bit: the weights (23M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run paraphrase-MiniLM-L6-v2 on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use paraphrase-MiniLM-L6-v2 commercially?

Yes. paraphrase-MiniLM-L6-v2 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

What is paraphrase-MiniLM-L6-v2's context length?

512 tokens, from the maximum position embeddings in its published configuration.

Similar Models

Model · Sentence similarity

all-MiniLM-L6-v2

Sentence Transformers

This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. Using this model becomes easy when you have sentence-transformers installed: Then you can use the model like this: Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. The project aims to train sentence embedding models on very large sentence level datasets using a self-supervised contrastive learning objective. We used the pretrained…

Open weights apache-2.0 23M parameters 512 tokens sentence-transformers

View model

Model · Sentence similarity

multi-qa-MiniLM-L6-cos-v1

Sentence Transformers

This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and was designed for semantic search. It has been trained on 215M (question, answer) pairs from diverse sources. For an introduction to semantic search, have a look at: SBERT.net - Semantic Search Using this model becomes easy when you have sentence-transformers installed: Then you can use the model like this: Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the correct pooling-operation on-top of the contextualized word embeddings. Similarly to the PyTorch example above, to use the model…

Open weights 23M parameters 512 tokens sentence-transformers

View model

Model · Sentence similarity

paraphrase-MiniLM-L3-v2

Sentence Transformers

This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. Using this model becomes easy when you have sentence-transformers installed: Then you can use the model like this: Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. This model was trained by sentence-transformers. If you find this model helpful, feel free to cite our publication Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Open weights apache-2.0 17M parameters 512 tokens sentence-transformers

View model

Model · Sentence similarity

bge-micro-v2

Taylor

This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. Distilled in a 2-step training process (bge-micro was step 1) from BAAI/bge-small-en-v1.5. Using this model becomes easy when you have sentence-transformers installed: Then you can use the model like this: Without sentence-transformers, you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. For an automated evaluation of this model, see the Sentence Embeddings Benchmark: https://seb.sbert.net

Open weights mit 17M parameters 512 tokens sentence-transformers

View model

Model · Sentence similarity

rubert-tiny2

David Dale

This is an updated version of cointegrated/rubert-tiny: a small Russian BERT-based encoder with high-quality sentence embeddings. This post in Russian gives more details. The differences from the previous version include: - sentence embeddings approximate LaBSE closer than before; - meaningful segment embeddings (tuned on the NLI task) - the model is focused only on Russian. The model should be used as is to produce sentence embeddings (e.g. for KNN classification of short texts) or fine-tuned for a downstream task. Sentence embeddings can be produced as follows: Alternatively, you can use the model with sentencetransformers: For those who want to run the inference with VLLM, there is a…

Open weights mit 29M parameters 2,048 tokens sentence-transformers

View model

Model · Sentence similarity

snowflake-arctic-embed-s

Snowflake

12/04/2024: Release of snowflake-arctic-embed-l-v2.0 and snowflake-arctic-embed-m-v2.0 our newest models with multilingual workloads in mind. These models outperform prior versions of Arctic Embed and we suggest these replace prior versions! 07/26/2024: Release preprint [[2407.18887] Embedding And Clustering Your Data Can Improve Contrastive Pretraining](https://arxiv.org/abs/2407.18887) on arXiv. 07/18/2024: Release of snowflake-arctic-embed-m-v1.5, capable of producing highly compressible embedding vectors that preserve quality even when squished as small as 128 bytes per vector. Details about the development of this model are available in the launch post on the Snowflake engineering…

Open weights apache-2.0 33M parameters 512 tokens sentence-transformers

View model