bge-small-en-v1.5

BAAI MIT out 3 years ago

33.4M parameters, text in, embeddings out, 512 tokens of context. 1 provider serves it.

Verified 2026-10-25 Rank #8 in Embeddings · score 108.1 method v1 Sources 3 Downloads 30d 54.1M · #1 of 619 tracked

Changed this week: 2 new quantized builds on the Hub (33 in all).2026-10-24

Can I use it?MITPermissivelicence file2026-10-24

Does it work?64.3on MTEB#16 of 64 measuredleaderboard2026-10-24

Do people use it?54.1Mdownloads, 30 days#1 of 619 we track · top 1%+8 likes, downloads −7% this week651 likesHugging Face2026-10-24

Worth reading

Picked by rule from 4 tracked pieces: the lab's own post, the most shared, a field report, the latest.

4 best of 4 tracked

The 1 pick above, plus these 3. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

Articlebge-small-en-v1.5 huggingface.co api & BAAI bge-small-en-v1.5 github AI Model - Toolifytoolify.ai · undated
ArticleBGE Small EN v1.5 by BAAI | The GTM Directorythegtmdirectory.com · undated
ArticleBGE Small EN v1.5 — Specs, Pricing & GPU Requirements | InferenceBench | InferenceBenchinferencebench.io · undated

Where to run it

Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.

#	Cheapest live row	In /1M	Out /1M	Context	Quant	Seen
1	HF Inference	—	—	—	—	Oct 25

All 1 provider →

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

Served by 1 API provider; 33 quantized builds for local use; a big base for builders (554 fine-tunes, adapters and merges).

As a service1HF Inference
On own hardware33quantized builds
Built on554551 fine-tunes · 3 adapters · 0 merges
In apps100+Hugging Face Spaces

Top Spaces running it: embedl/hfviewer ♥ 181 · Sarath0x8f/Document-QA-bot ♥ 33 · AI-ANK/RAGArch ♥ 19 · fau/videoxity ♥ 13

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Fits a 16 GB card at BF16

About 0.1 GB at BF16: weights only, so leave room for the KV cache.

Estimate params × bytes per param × 1.2 · how we estimate

How it scores

Third-party measurements only; author-reported numbers stay in Reference, marked as such.

#16 of 64 on MTEB so far

64.3. Each tick is a model on this site with a measured MTEB row; author-reported numbers stay in Reference. Mean(task) x100 over 41 tasks. 41 tasks, split=test

Source · leaderboard · 2026-10-24

The story so far

Weights, every dated article, and repo commits on one rail.

14 repo commits since the weights landed

Teal is the weights (2023-09-12), grey ticks are repo commits.

What people built on it

587 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

587 derivatives on the Hub, most of them fine-tunes

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

GGUF ChristianAzinn/bge-small-en-v1.5-gguf for llama.cpp, Ollama, LM Studio 81.6k downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Benchmarks

3 rows

Benchmarks 3 rows

measured third-party leaderboard row (3)author-reported number from the model card or eval filesNo confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.

Benchmark	Value	Setting / notes	Reported by	Date
measured MTEB · Embed text for search	53.86 nDCG@10 x100 (mean of 10 tasks)	10 tasks, split=test	leaderboard	2026-10-24
measured MTEB · Embed text for search	64.3 mean(task) x100 over 41 tasks	41 tasks, split=test	leaderboard	2026-10-24
measured MTEB Reranking · Rerank search results	46.94 MAP x100 (mean of 2 tasks)	2 tasks, split=test	leaderboard	2026-10-24

Use-case evidence

2 use cases

What it is good at 2 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Facts

21 fields

Listed in Embeddings

Facts

bge-small-en-v1.5 has 33.4M parameters. (safetensors metadata, observed 2026-10-24)
Context length is 512 tokens (512). (config.json max_position_embeddings, observed 2026-10-24)
License: MIT (friction: none). Permissive. (license file, observed 2026-10-24)
Created 2023-09-12, last modified 2024-02-22. (HF API, observed 2026-10-24)

Pipeline tag	feature-extraction
Library	sentence-transformers
Architectures	BertModel
Model type	bert
Dtype	float32
Params by dtype	F32: 33,360,000 · I64: 512
Size band	S
MoE	no
Context length	512 (config.json max_position_embeddings)
Languages	en
Papers	arXiv:2401.03462, arXiv:2312.15503, arXiv:2311.13534, arXiv:2310.07554, arXiv:2309.07597
License	MIT (mit)
Hub snapshot	2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · BertModel · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · BertModel · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for bge-small-en-v1.5: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)	0.1 GB · Edge (<4 GB)
FP8 (1 byte/param)	0.0 GB · Edge (<4 GB)
INT4 (0.5 bytes/param)	0.0 GB · Edge (<4 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.

All variants

40 variants

Lineage

Original model Family: BAAI/bge-small-v1.5 587 derivatives on the Hub

quantized: 33finetune: 551adapter: 3

Top variants (40 shown, ranked by downloads)

Variant	Relation	Quant	Params	Band	License	Author	Likes	Downloads
Xenova/bge-small-en-v1.5	quantized	—	—	—	—	Xenova	20	3.5M
farbodtavakkoli/OTel-Embedding-33M	finetune	—	33.4M	S	apache-2.0	farbodtavakkoli	0	1.2M
abhinand/MedEmbed-small-v0.1	finetune	—	33.4M	S	apache-2.0	abhinand	24	1.2M
darkmaniac7/bge-small-en-MNN	quantized	—	—	—	mit	darkmaniac7	0	90.4k
ChristianAzinn/bge-small-en-v1.5-gguf	quantized	gguf	—	—	mit	ChristianAzinn	2	81.6k
tomaarsen/setfit-absa-bge-small-en-v1.5-restaurants-aspect	finetune	—	—	—	apache-2.0	tomaarsen	4	48.3k
tomaarsen/setfit-absa-bge-small-en-v1.5-restaurants-polarity	finetune	—	—	—	apache-2.0	tomaarsen	0	48.4k
baconnier/Finance2_embedding_small_en-V1.5	finetune	—	33.4M	S	—	baconnier	1	37.9k
kasparas12/is_organizational_model	finetune	—	33.4M	S	—	kasparas12	0	27k
kasparas12/crypto_individual_infer_model_setfit	finetune	—	—	—	—	kasparas12	3	24.4k
kasparas12/crypto_organization_infer_model_setfit	finetune	—	33.4M	S	—	kasparas12	3	24.4k
ggml-org/bge-small-en-v1.5-Q8_0-GGUF	quantized	Q8_0	—	—	mit	ggml-org	6	21.5k
austinpatrickm/finetuned_bge_embeddings_v5_small_v1.5	finetune	—	33.4M	S	—	austinpatrickm	0	19.9k
onnx-community/bge-small-en-v1.5-ONNX	quantized	—	—	—	—	onnx-community	0	10.9k
danielkorat/bge-small-en-v1.5_setfit-sst2-english	finetune	—	33.4M	S	—	danielkorat	0	10.1k
SalmanFaroz/DisEmbed-v1	finetune	—	33.4M	S	mit	SalmanFaroz	0	5.3k
Bombek1/bge-small-en-v1.5-litert	finetune	—	—	—	mit	Bombek1	0	3.5k
prithivida/bge-base-en-v1.5-gguf	quantized	gguf	—	—	mit	prithivida	1	3.1k
austinpatrickm/finetuned_bge_embeddings_v6_small_v1.5	finetune	—	33.4M	S	—	austinpatrickm	0	3.1k
waghelad/ds-rag-embedder-v1	finetune	—	33.4M	S	apache-2.0	waghelad	0	3.1k
raul-delarosa99/bge-small-en-v1.5-RIRAG_ObliQA	finetune	—	33.4M	S	—	raul-delarosa99	0	3k
knowledgator/modern-gliner-bi-base-v1.0	finetune	—	—	—	apache-2.0	knowledgator	27	1.6k
JordanTallon/Unifeed	finetune	—	33.4M	S	—	JordanTallon	1	2.5k
sartifyllc/MultiLinguSwahili-bge-small-en-v1.5-nli-matryoshka	finetune	—	33.4M	S	—	sartifyllc	1	2.1k
RiverRider/motherlode-code-small-en-v0.1	finetune	GGUF	33.4M	S	other	RiverRider	3	1.5k
albertobarnabo/ecommerce-product-search-embeddings	quantized	—	33.4M	S	apache-2.0	albertobarnabo	1	1.2k
elsayovita/bge-small-en-v1.5-esg	finetune	—	33.4M	S	apache-2.0	elsayovita	0	1.1k
cstr/bge-small-en-v1.5-GGUF	quantized	GGUF	—	—	mit	cstr	0	1.1k
archisgore/bge-small-winetone	finetune	—	33.4M	S	—	archisgore	0	977
wangjinzzhong/bge-small-en-v1.5-Q4_K_M-GGUF	quantized	GGUF	—	—	mit	wangjinzzhong	1	852
ArchitRastogi/BGE-Small-LegalEmbeddings-USCode	finetune	—	33.4M	S	apache-2.0	ArchitRastogi	1	848
Stffens/bge-small-rrf-lme-v1	finetune	—	33.4M	S	mit	Stffens	1	784
Stffens/bge-small-rrf-v2	quantized	—	33.4M	S	apache-2.0	Stffens	1	748
vivekkopthsd/fiqa-retriever-bge-small	finetune	—	33.4M	S	—	vivekkopthsd	0	767
Stffens/bge-small-rrf-v3	finetune	—	33.4M	S	mit	Stffens	0	743
rohithbojja/intent-classification-small	finetune	—	33.4M	S	—	rohithbojja	0	739
Ihor/gliner-biomed-bi-small-v1.0	finetune	—	—	—	apache-2.0	Ihor	3	588
SriRamanaAtmic/AtmicQuoterv1	finetune	—	33.4M	S	—	SriRamanaAtmic	0	668
Huzaifa68/investment_v1	finetune	—	33.4M	S	—	Huzaifa68	1	611
himanshu23099/bge_embedding_finetune1	finetune	—	33.4M	S	—	himanshu23099	0	599

Repository changelog

14 commits

changed in the last 7 daysOther: 2026-10-24all changes

Repository changelog 14 commits

Date	Commit	SHA
2024-02-22	onnx support (#9)	5c38ec7
2024-02-02	Update README.md	35baf77
2023-12-18	Adding `safetensors` variant of this model (#3)	c202d20
2023-11-02	Update tokenizer_config.json (#5)	9825324
2023-10-12	update	b49342c
2023-09-27	Update modules.json	8888563
2023-09-15	update	73f5855
2023-09-14	Update README.md	ceaf083
2023-09-12	update	6aad939
2023-09-12	update	8120161
2023-09-12	update	45e16d1
2023-09-12	update	0b329b7
2023-09-12	upload	9b09f79
2023-09-12	initial commit	a32952c

All providers

1 provider · 1 endpoint

Where to run 1 providers · 1 endpoints

Hosted endpoints observed for bge-small-en-v1.5, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting.

Provider	Model ref	Context	Quant	$/M in	$/M out	tok/s	TTFT	Tools	Status	Observed
HF Inference hf	BAAI/bge-small-en-v1.5	—	—	—	—	—	—	—	live	2026-10-25

Articles and coverage

4 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.

Analysis4

bge-small-en-v1.5 by BAAI - Overview, Benchmarks & Hardware | Q4KM.ai

q4km.ai895 words

bge-small-en-v1.5 is a lightweight English sentence-embedding model released by BAAI that maps text to a dense 384-dimensional vector. The model features a 12-layer distilled BERT backbone and fits comfortably in 2 GB of GPU memory for inference. It achieves a latency of approximately 2-3 ms per sentence on an RTX 3060 and is likely released under the MIT License.written, reviewed, cited

A modern 8-core CPU can run the model at approximately 200-300 tokens per second.
The repository lists the license as unknown, though the tag license:mit suggests an MIT License.

bge-small-en-v1.5 huggingface.co api & BAAI bge-small-en-v1.5 github AI Model - Toolify

toolify.ai

BGE Small EN v1.5 by BAAI | The GTM Directory

thegtmdirectory.com1,465 words

BAAI released BGE Small EN v1.5, a compact open-weight English embedding model producing 384-dimensional vectors under tight resource constraints. It costs $0.02 per million tokens on Cloudflare Workers AI and supports a 512-token max input per passage with English-only support. The model achieved an MTEB (English v2) score of 64.3 and carries an MIT license for local or air-gapped deployments.written, reviewed, cited

The v1.5 update removes the need for query instruction prefixes and alleviates similarity distribution problems found in earlier releases.

BGE Small EN v1.5 — Specs, Pricing & GPU Requirements | InferenceBench | InferenceBench

inferencebench.io441 words

BGE Small EN v1.5 is a 0.033B parameter dense model from BAAI featuring Multi-Head Attention with 12 layers and 384 hidden dimensions. It requires approximately 0.1 GB of VRAM at BF16 precision and achieves 3.5K tokens/sec throughput on a B200 SXM GPU. Inference costs $4261 per month, equating to $0.46 per million tokens.written, reviewed, cited

Papers

5 papers

Papers 5

arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).

Long Context Compression with Activation Beacon cited in card

Peitian Zhang, Zheng Liu, Shitao Xiao +3 · 2024-01-07 · PDF · HF paper page (30 upvotes)
A Multi-Task Embedder For Retrieval Augmented LLMs cited in card

Peitian Zhang, Shitao Xiao, Zheng Liu +2 · 2023-10-11 · PDF · HF paper page (6 upvotes)
LM-Cocktail: Resilient Tuning of Language Models via Model Merging cited in card

Shitao Xiao, Zheng Liu, Peitian Zhang +1 · 2023-11-22 · PDF · HF paper page (3 upvotes)
Llama2Vec: Unsupervised Adaptation of Large Language Models for Dense Retrieval cited in card

Zheng Liu, Chaofan Li, Shitao Xiao +2 · 2023-12-24 · PDF · HF paper page (2 upvotes)
C-Pack: Packed Resources For General Chinese Embeddings cited in card

Shitao Xiao, Zheng Liu, Peitian Zhang +3 · 2023-09-14 · PDF · HF paper page (2 upvotes)

Similar models

3 models

Similar models

jina-embeddings-v2-small-en · 32.7M
same size band (S), same task (feature-extraction), permissive license
bge-small-zh-v1.5 · 24.0M
same size band (S), same task (feature-extraction), permissive license
granite-embedding-small-english-r2 · 47.7M
same size band (S), same task (feature-extraction), permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:32 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:00 UTC · Last verified 2026-10-25

Follow bge-small-en-v1.5

Watch this model

/feeds/models/BAAI--bge-small-en-v1.5.xml(RSS 2.0) · every change to bge-small-en-v1.5
/feeds/families/BAAI--bge-small-v1.5.xml(RSS 2.0) · every change across the bge-small-v1.5 family

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.