SAVRN
Search Contact SAVRN

Open-weight model · Sentence similarity

multilingual-e5-small

by Liang Wang intfloat/multilingual-e5-small

Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei, arXiv 2024 This model has 12 layers and the embedding size is 384. Below is an example to encode queries and passages from the MS-MARCO passage ranking dataset.

Parameters118M
Context512
Weights2.2 GB
Licensemit
AccessOpen weights
Monthly Downloads12.3M

Runs On

What it takes to serve multilingual-e5-small (118M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

PrecisionWeightsMemory neededCheapest setupPer hourAlso fits
16-bit 0.2 GB 0.3 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
8-bit 0.1 GB 0.1 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
4-bit 0.1 GB 0.1 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 multilingual-e5-small

Two numbers on this page seem to disagree: 2.28 GB across 43 files on disk, and 0.3 GB of memory at 16-bit. The repository carries the 118-million-parameter encoder in four formats, safetensors, ONNX, OpenVINO and PyTorch, at float32. Running, it needs 0.3 GB, or 0.1 GB at 8-bit or 4-bit, so the Index's cheapest setup, one MI300X with 192 GB at $1.85 an hour, sits nearly empty. Its job is sentence similarity, with 12 layers, 384-dimensional embeddings and 100 languages behind it.

The license could hardly be shorter: MIT permits commercial use, modification and redistribution with only the copyright and permission notices kept, so a search index built on it carries no license strings. The 512-token context caps each passage you encode, so long documents get split. The publisher says it was initialized from Multilingual-MiniLM-L12-H384 and warns that low-resource languages may see degradation, so test your languages first.

Model Card

By Liang Wang, published under mit, revision 614241f622f5.

Multilingual E5 Text Embeddings: A Technical Report. Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei, arXiv 2024

This model has 12 layers and the embedding size is 384.

Usage

Below is an example to encode queries and passages from the MS-MARCO passage ranking dataset.

Read the full model card (900 words)

Configuration

Architecture
BertModel
Context length (tokens)
512
Layers
12
Hidden size
384
Feed-forward size
1,536
Attention heads
12
Vocabulary size
250,037
Stored precision
float32
Model type
bert

Identity and Version

Repository
intfloat/multilingual-e5-small
Publisher
Liang Wang
Task
Sentence similarity
Modality
Text
Library
sentence-transformers
Parameters
118M parameters
Languages
af, am, ar, as, az, be, bg, bn
Revision
614241f622f53c4eeff9890bdc4f31cfecc418b3
First published
2023-06-30
Last updated
2026-04-02

Files and Weights

43 files, 2.3 GB in total. The weights are 6 files totalling 2.2 GB in bin, onnx, safetensors.

Weights6 files · 2.2 GB
Configuration28 files · 16.0 KB
Tokenizer4 files · 34.2 MB
Documentation1 file · 497.5 KB
Other3 files · 10.5 MB
Repository1 file · 1.6 KB
Every file
FileTypeSizeSHA-256
model.safetensorsWeights470.6 MB 1a55775f5344
onnx/model.onnxWeights470.3 MB ca456c06b3a9
onnx/model_O4.onnxWeights235.1 MB 4654c156f3e4
onnx/model_qint8_avx512_vnni.onnxWeights118.3 MB dd476dd0c251
openvino/openvino_model.binWeights470.0 MB 22f5a10ec005
pytorch_model.binWeights470.7 MB f0ccb64a2edc
.eval_results/ArguAna.yamlConfiguration575 B
.eval_results/BrightAopsRetrieval.yamlConfiguration641 B
.eval_results/BrightBiologyLongRetrieval.yamlConfiguration651 B
.eval_results/BrightBiologyRetrieval.yamlConfiguration647 B
.eval_results/BrightEarthScienceLongRetrieval.yamlConfiguration663 B
.eval_results/BrightEarthScienceRetrieval.yamlConfiguration659 B
.eval_results/BrightEconomicsLongRetrieval.yamlConfiguration655 B
.eval_results/BrightEconomicsRetrieval.yamlConfiguration651 B
.eval_results/BrightLeetcodeRetrieval.yamlConfiguration651 B
.eval_results/BrightPonyLongRetrieval.yamlConfiguration645 B
.eval_results/BrightPonyRetrieval.yamlConfiguration641 B
.eval_results/BrightPsychologyLongRetrieval.yamlConfiguration655 B
.eval_results/BrightPsychologyRetrieval.yamlConfiguration653 B
.eval_results/BrightRoboticsLongRetrieval.yamlConfiguration653 B
.eval_results/BrightRoboticsRetrieval.yamlConfiguration647 B
.eval_results/BrightStackoverflowLongRetrieval.yamlConfiguration665 B
.eval_results/BrightStackoverflowRetrieval.yamlConfiguration659 B
.eval_results/BrightSustainableLivingLongRetrieval.yamlConfiguration673 B
.eval_results/BrightSustainableLivingRetrieval.yamlConfiguration667 B
.eval_results/BrightTheoremQAQuestionsRetrieval.yamlConfiguration669 B
.eval_results/BrightTheoremQATheoremsRetrieval.yamlConfiguration667 B
1_Pooling/config.jsonConfiguration200 B
config.jsonConfiguration655 B
modules.jsonConfiguration387 B
onnx/config.jsonConfiguration653 B
onnx/special_tokens_map.jsonConfiguration167 B
sentence_bert_config.jsonConfiguration57 B
special_tokens_map.jsonConfiguration167 B
README.mdDocumentation497.5 KB
onnx/sentencepiece.bpe.modelOther5.1 MB cfc8146abe2a
openvino/openvino_model.xmlOther363.4 KB
sentencepiece.bpe.modelOther5.1 MB cfc8146abe2a
.gitattributesRepository1.6 KB
onnx/tokenizer.jsonTokenizer17.1 MB 0b44a9d7b51c
onnx/tokenizer_config.jsonTokenizer443 B
tokenizer.jsonTokenizer17.1 MB 0b44a9d7b51c
tokenizer_config.jsonTokenizer443 B

License and Download

License
mit
Access
Open weights, no gate
Download size
2.2 GB
Download from Liang Wang

Released by Liang Wang through its official repository on Hugging Face. Read the license.

Built From

Evaluations

Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical.

BenchmarkConditionsResultReported byRevisionDate
mteb/BRIGHT Task BrightAopsRetrievalMetric BrightAopsRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 7.921 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightAopsRetrieval_default_standardMetric BrightAopsRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 7.921 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightBiologyLongRetrievalMetric BrightBiologyLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.709 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightBiologyLongRetrieval_default_longMetric BrightBiologyLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.709 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightBiologyRetrievalMetric BrightBiologyRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 4.237 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightBiologyRetrieval_default_standardMetric BrightBiologyRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 4.237 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEarthScienceLongRetrievalMetric BrightEarthScienceLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 14.943 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEarthScienceLongRetrieval_default_longMetric BrightEarthScienceLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 14.943 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEarthScienceRetrievalMetric BrightEarthScienceRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 19.293 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEarthScienceRetrieval_default_standardMetric BrightEarthScienceRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 19.293 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEconomicsLongRetrievalMetric BrightEconomicsLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 10.68 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEconomicsLongRetrieval_default_longMetric BrightEconomicsLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 10.68 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEconomicsRetrievalMetric BrightEconomicsRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 7.414 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightEconomicsRetrieval_default_standardMetric BrightEconomicsRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 7.414 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightLeetcodeRetrievalMetric BrightLeetcodeRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 23.579 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightLeetcodeRetrieval_default_standardMetric BrightLeetcodeRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 23.579 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPonyLongRetrievalMetric BrightPonyLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 0.749 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPonyLongRetrieval_default_longMetric BrightPonyLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 0.749 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPonyRetrievalMetric BrightPonyRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 0.184 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPonyRetrieval_default_standardMetric BrightPonyRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 0.184 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPsychologyLongRetrievalMetric BrightPsychologyLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 2.97 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPsychologyLongRetrieval_default_longMetric BrightPsychologyLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 2.97 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPsychologyRetrievalMetric BrightPsychologyRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 7.909 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightPsychologyRetrieval_default_standardMetric BrightPsychologyRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 7.909 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightRoboticsLongRetrievalMetric BrightRoboticsLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 8.911 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightRoboticsLongRetrieval_default_longMetric BrightRoboticsLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 8.911 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightRoboticsRetrievalMetric BrightRoboticsRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.57 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightRoboticsRetrieval_default_standardMetric BrightRoboticsRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.57 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightStackoverflowLongRetrievalMetric BrightStackoverflowLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 10.256 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightStackoverflowLongRetrieval_default_longMetric BrightStackoverflowLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 10.256 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightStackoverflowRetrievalMetric BrightStackoverflowRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 6.857 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightStackoverflowRetrieval_default_standardMetric BrightStackoverflowRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 6.857 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightSustainableLivingLongRetrievalMetric BrightSustainableLivingLongRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 10.417 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightSustainableLivingLongRetrieval_default_longMetric BrightSustainableLivingLongRetrieval_default_longSetup Obtained using MTEB v2.10.12Comparison conditions not established 10.417 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightSustainableLivingRetrievalMetric BrightSustainableLivingRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.101 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightSustainableLivingRetrieval_default_standardMetric BrightSustainableLivingRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.101 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightTheoremQAQuestionsRetrievalMetric BrightTheoremQAQuestionsRetrievalSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.877 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/BRIGHT Task BrightTheoremQAQuestionsRetrieval_default_standardMetric BrightTheoremQAQuestionsRetrieval_default_standardSetup Obtained using MTEB v2.10.12Comparison conditions not established 9.877 Obtained using MTEB v2.10.12
Reported by a third party
Evaluated revision not stated 2026-03-31
mteb/arguana Task ArguAnaMetric ArguAnaSetup Obtained using MTEB v1.36.26Comparison conditions not established 39.088 Obtained using MTEB v1.36.26
Reported by a third party
Evaluated revision not stated 2026-04-02
mteb/arguana Task ArguAna_default_testMetric ArguAna_default_testSetup Obtained using MTEB v1.36.26Comparison conditions not established 39.088 Obtained using MTEB v1.36.26
Reported by a third party
Evaluated revision not stated 2026-04-02

Memory Requirements

PrecisionWeights in memory
As published2.2 GB
16-bit0.2 GB
8-bit0.1 GB
4-bit0.1 GB

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

Compare multilingual-e5-small

Questions About multilingual-e5-small

How much GPU memory does multilingual-e5-small need?

About 0.3 GB at 16-bit and 0.1 GB at 4-bit: the weights (118M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run multilingual-e5-small 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 multilingual-e5-small commercially?

Yes. multilingual-e5-small is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.

What is multilingual-e5-small's context length?

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

Similar Models

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 118M parameters 512 tokens sentence-transformers

Model · Sentence similarity

contact_dedupe_STx_model

Gourav

This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…

Open weights 118M parameters 512 tokens sentence-transformers

Model · Sentence similarity

ko-sroberta-multitask

Junghyun Gan

This is a sentence-transformers model: It maps sentences & paragraphs to a 768 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. KorSTS, KorNLI 학습 데이터셋으로 멀티 태스크 학습을 진행한 후 KorSTS 평가 데이터셋으로 평가한 결과입니다. The model was trained with the parameters: sentencetransformers.datasets.NoDuplicatesDataLoader.NoDuplicatesDataLoader of…

Open weights 111M parameters 514 tokens sentence-transformers

Model · Sentence similarity

lt-un-data-fine-fine-es

Dell Research Harvard

This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…

Open weights 110M parameters 512 tokens sentence-transformers

Model · Sentence similarity

lt-wikidata-comp-es

Dell Research Harvard

This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…

Open weights 110M parameters 512 tokens sentence-transformers

Model · Sentence similarity

lt-un-data-fine-coarse-es

Dell Research Harvard

This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we…

Open weights 110M parameters 512 tokens sentence-transformers