gte-large

thenlper MIT out 3 years ago

335M parameters, text in, embeddings out, 512 tokens of context. 2 providers serve it from $0.010 per million input tokens.

Verified 2026-10-25 Rank #59 in Embeddings · score 86.5 method v1 Sources 3 Downloads 30d 564k · #157 of 619 tracked

Can I run it?0.2GB at INT4Estimatefits a 16 GB cardmethod

What does it cost?$0.010/1M incheapest of 1DeepInfra2026-10-24

Can I use it?MITPermissivelicence file2026-10-24

Does it work?64.77on MTEB#15 of 64 measuredleaderboard2026-10-24

Do people use it?564kdownloads, 30 days#157 of 619 we track · top 25%+1 likes, downloads +15% this week308 likesHugging Face2026-10-24

Worth reading

Picked by rule from 2 tracked pieces: the lab's own post, the most shared, a field report, the latest.

2 best of 2 tracked

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

ArticleGeneral Text Embeddings (GTE) English (gte_large) | gte_large | Spark NLP 5.0.2sparknlp.org · 2023-08-15

Where to run it

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

#	Cheapest 2	In /1M	Out /1M	Context	Quant	Seen
1	DeepInfra	$0.010	—	512	—	Oct 24
2	HF Inference	—	—	—	—	Oct 25

All 2 providers →

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 2 API providers; 7 quantized builds for local use.

As a service2DeepInfra, HF Inference
On own hardware7quantized builds
Built on2929 fine-tunes · 0 adapters · 0 merges
In apps100+Hugging Face Spaces

Top Spaces running it: AI-ANK/RAGArch ♥ 19 · captain-awesome/Chat_with_any_website ♥ 19 · amaai-lab/SonicVerse ♥ 13 · Karthikeyen92/Stock-Sentiment-Analysis ♥ 6

Will it fit?

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

Fits a 16 GB card at INT4

About 0.2 GB at INT4, 0.4 GB at FP8, 0.8 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.

#15 of 64 on MTEB so far

64.77. 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.

1 dated article and 24 repo commits so far

Teal is the weights (2023-07-27), amber triangles are 1 dated article (each links to it), grey ticks are repo commits.

What people built on it

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

36 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/gte-large-gguf for llama.cpp, Ollama, LM Studio 26.9k downloads

All variants →

Reference

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

Benchmarks

11 rows

Benchmarks 11 rows

measured third-party leaderboard row (3)author-reported number from the model card or eval files (8)No 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	53.29 nDCG@10 x100 (mean of 10 tasks)	10 tasks, split=test	leaderboard	2026-10-24
measured MTEB · Embed text for search	64.77 mean(task) x100 over 41 tasks	41 tasks, split=test	leaderboard	2026-10-24
measured MTEB Reranking · Rerank search results	47.84 MAP x100 (mean of 2 tasks)	2 tasks, split=test	leaderboard	2026-10-24
author-reported MTEB · Embed text for search	63.13	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	46.84	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	85	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	59.13	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	52.22	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	83.35	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	31.66	gte-large	model author	2026-10-07
author-reported MTEB · Embed text for search	73.33	gte-large	model author	2026-10-07

What the card says

1 cited sentence

Summary generated, cited

It is released under an MIT license and is used for applications like information retrieval and semantic textual similarity.^[1]^[2]

license
“information retrieval, semantic textual similarity”

Generated 2026-10-07 15:04 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.

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

gte-large has 335M 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 2023-07-27, last modified 2024-11-15. (HF API, observed 2026-10-24)

Pipeline tag	sentence-similarity
Library	sentence-transformers
Architectures	BertModel
Model type	bert
Dtype	float16
Params by dtype	I64: 512 · F16: 335,141,888
Size band	S
MoE	no
Context length	512 (config.json max_position_embeddings)
Languages	en
Papers	arXiv:2308.03281
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 gte-large: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)	0.8 GB · Edge (<4 GB)
FP8 (1 byte/param)	0.4 GB · Edge (<4 GB)
INT4 (0.5 bytes/param)	0.2 GB · Edge (<4 GB)

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

All variants

36 variants

Lineage

Original model Family: thenlper/gte-large 36 derivatives on the Hub

quantized: 7finetune: 29

Top variants (36 shown, ranked by downloads)

Variant	Relation	Quant	Params	Band	License	Author	Likes	Downloads
ChristianAzinn/gte-large-gguf	quantized	gguf	—	—	mit	ChristianAzinn	1	26.9k
mradermacher/gte-large-i1-GGUF	quantized	GGUF	—	—	mit	mradermacher	0	4.7k
Xenova/gte-large	quantized	—	—	—	—	Xenova	2	3.3k
mradermacher/gte-large-GGUF	quantized	GGUF	—	—	mit	mradermacher	0	2.6k
algolia/algolia-large-en-generic-v2410	quantized	—	335M	S	mit	algolia	0	2.1k
mjwong/gte-large-mnli-anli	finetune	—	335M	S	cc-by-nc-4.0	mjwong	1	898
richtext/foodmapper-gte-large	finetune	—	—	—	mit	richtext	0	400
JFernandoGRE/gtelarge-colombian-elitenames	finetune	—	335M	S	—	JFernandoGRE	0	392
groonga/gte-large-Q4_K_M-GGUF	quantized	GGUF	—	—	mit	groonga	0	312
JFernandoGRE/gtelarge-colombian-corporationnames	finetune	—	335M	S	—	JFernandoGRE	0	311
Ramyashree/gte-large-train-test-4	finetune	—	335M	S	—	Ramyashree	1	258
Ramyashree/gte-large-train-test-2	finetune	—	335M	S	—	Ramyashree	0	226
Ramyashree/gte-large-train-test-3	finetune	—	335M	S	—	Ramyashree	0	224
Ramyashree/gte-large-with500records-validate	finetune	—	335M	S	—	Ramyashree	0	210
Ramyashree/gte-large-train-test	finetune	—	335M	S	—	Ramyashree	0	206
HarshalBhg/gte-large-setfit-train-test2	finetune	—	335M	S	—	HarshalBhg	0	205
JFernandoGRE/gtelarge-duplicates-judgenames	finetune	—	335M	S	apache-2.0	JFernandoGRE	0	135
HarshalBhg/gte-large-setfit-train-b77-test3	finetune	—	335M	S	—	HarshalBhg	0	57
llmware/gte-large-onnx	quantized	—	—	—	apache-2.0	llmware	0	46
jimnoneill/gte-large-static-256	finetune	—	7.59M	S	mit	jimnoneill	0	39
Ramyashree/gte-large-with80records	finetune	—	335M	S	—	Ramyashree	0	38
devrim/reign-base-l3_gn-gte-large_dapfam-ftwarm-c512s384	finetune	—	22.6M	S	apache-2.0	devrim	0	33
Ramyashree/gte-large-with500records	finetune	—	335M	S	—	Ramyashree	0	32
HarshalBhg/gte-large-setfit-train-test1	finetune	—	335M	S	—	HarshalBhg	0	32
devrim/reign-small-l2_gn-gte-large_st-256_val-selected	finetune	—	4.09M	S	apache-2.0	devrim	0	31
Ramyashree/gte-large-with500records-test	finetune	—	335M	S	—	Ramyashree	0	30
devrim/reign-small-l2_gn-gte-large_st-384_val-selected	finetune	—	4.09M	S	apache-2.0	devrim	0	30
neih4207/checkpoint	finetune	—	335M	S	mit	neih4207	0	29
devrim/reign-small-l2_gn-gte-large_st-128_val-selected	finetune	—	4.09M	S	apache-2.0	devrim	0	29
devrim/reign-base-l3_gn-gte-large_dapfam-ftwarm-c512s512	finetune	—	22.6M	S	apache-2.0	devrim	0	28
devrim/reign-small-l2_gn-gte-large_st-512_val-selected	finetune	—	4.09M	S	apache-2.0	devrim	0	26
devrim/reign-tiny-l1_gn-gte-large_val-selected	finetune	—	681K	S	apache-2.0	devrim	0	25
devrim/reign-base-l3_gn-gte-large_val-selected	finetune	—	22.6M	S	apache-2.0	devrim	0	23
devrim/reign-small-l2_gn-gte-large_val-selected	finetune	—	4.09M	S	apache-2.0	devrim	0	23
devrim/reign-large-l4_gn-gte-large_s384_val-selected	finetune	—	52.5M	S	apache-2.0	devrim	0	23
kareem2222/MusicalChatBott	finetune	—	—	—	mit	kareem2222	0	0

Repository changelog

24 commits

Repository changelog 24 commits

Date	Commit	SHA
2024-11-15	Add exported onnx model 'model_qint8_avx512_vnni.onnx' (#20)	4bef63f
2024-11-15	Add exported onnx model 'model_O4.onnx' (#21)	6356d99
2024-11-15	Add exported openvino model 'openvino_model.xml' (#22)	170fce7
2024-11-15	Add exported openvino model 'openvino_model_qint8_quantized.xml' (#23)	ffde661
2024-02-05	Update README.md	5857861
2023-09-25	Adding `safetensors` variant of this model (#11)	8cb729e
2023-08-25	Adding ONNX file of this model (#6)	2b5163b
2023-08-15	Fix typo in model name (#8)	ef03ec3
2023-08-08	Update README.md	613856b
2023-08-08	Update README.md	5868460
2023-08-04	Upload modules.json	8345a94
2023-08-02	Update README.md	ccf2953
2023-08-01	Upload sentence_bert_config.json	bb4efcf
2023-08-01	Upload config.json	950114b
2023-08-01	Delete 1_pooling	39bc219
2023-08-01	Upload config.json	bedf446
2023-07-31	Update README.md	985f546
2023-07-31	add tags	c238a9f
2023-07-29	Delete training_args.bin	15d5d0a
2023-07-29	Update README.md	ec670d8
2023-07-27	Update README.md	ada1742
2023-07-27	Upload 6 files	97f4c0d
2023-07-27	Upload pytorch_model.bin	02a73bc
2023-07-27	initial commit	10fef62

All providers

2 providers · 2 endpoints

Where to run 2 providers · 2 endpoints

Hosted endpoints observed for gte-large, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.01/M.

Provider	Model ref	Context	Quant	$/M in	$/M out	tok/s	TTFT	Tools	Status	Observed
DeepInfra deepinfra	thenlper/gte-large	512	—	$0.01	—	—	—	—	live	2026-10-24
HF Inference hf	thenlper/gte-large	—	—	—	—	—	—	—	live	2026-10-25

Articles and coverage

2 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.

Analysis2

gte-large — text embeddings score and price · Undominated.ai

undominated.ai226 words

The gte-large model by Alibaba DAMO has no mean score on MTEB(eng, v2) because it completed only 21 of the benchmark's 41 tasks. DeepInfra and OpenRouter both charge $0.010 per 1M input tokens for this open weights model.generated, cited

General Text Embeddings (GTE) English (gte_large) | gte_large | Spark NLP 5.0.2

sparknlp.org2023-08-15583 words

Alibaba DAMO Academy released the General Text Embeddings (GTE) model family, including gte-large, gte-base, and gte-small variants built on the BERT framework. The gte-large model weighs 0.67 GB, features a 1024 dimension and 512 sequence length, and operates under an Open Source license. It requires Spark NLP 5.0.2+ compatibility and supports case-sensitive inputs for tasks like information retrieval and semantic textual similarity.written, reviewed, cited

Papers

1 paper

Papers 1

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).

Towards General Text Embeddings with Multi-stage Contrastive Learning cited in card

Zehan Li, Xin Zhang, Yanzhao Zhang +3 · 2023-08-07 · PDF · HF paper page (3 upvotes)

Similar models

3 models

Similar models

e5-large-v2 · 335M
same size band (S), same task (sentence-similarity), permissive license
embeddinggemma-300m · 303M
same size band (S), same task (sentence-similarity)
all-roberta-large-v1 · 355M
same size band (S), same task (sentence-similarity), fewer providers, permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:36 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:04 UTC · Last verified 2026-10-25

Follow gte-large

Watch this model

/feeds/models/thenlper--gte-large.xml(RSS 2.0) · every change to gte-large
/feeds/families/thenlper--gte-large.xml(RSS 2.0) · every change across the gte-large family

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