The story so far
Weights, every dated article, and repo commits on one rail.
17 repo commits since the weights landed
Teal is the weights (2023-06-30), grey ticks are repo commits.
multilingual-e5-large
560M parameters, text in, embeddings out, 514 tokens of context. 2 providers serve it from $0.010 per million input tokens.
Verified 2026-10-25 Rank #5 in Embeddings · score 111.1 method v1 Sources 3 Downloads 30d 6.5M · #20 of 619 tracked
Do people use it?6.5Mdownloads, 30 days#20 of 619 we track · top 3%+6 likes, downloads −3% this week1,281 likesHugging Face2026-10-24
Where to run it
Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.
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; heavily run locally (269 quantized builds); a big base for builders (197 fine-tunes, adapters and merges).
- As a service2DeepInfra, HF Inference
- On own hardware269quantized builds
- Built on197193 fine-tunes · 4 adapters · 0 merges
- In apps100+Hugging Face Spaces
Top Spaces running it: embedl/hfviewer ♥ 181 · nt3awnou/Nt3awnou-rescue-map ♥ 31 · antoinelouis/mtem-pruner ♥ 14 · HFforLegal/laws-retrieval ♥ 10
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 16 GB card at INT4
About 0.3 GB at INT4, 0.7 GB at FP8, 1.3 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.
#25 of 64 on MTEB so far
62.07. Each tick is a model on this site with a measured MTEB row; author-reported numbers stay in Reference. Mean over 40/41 tasks present. 40/41 tasks
Source · leaderboard · 2026-10-24
What people built on it
466 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
466 derivatives on the Hub, most of them quantized
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Reference
Everything else, folded: the full tables, every source.
Benchmarks
39 rows
Benchmarks 39 rows
measured third-party leaderboard row (3)author-reported number from the model card or eval files (36)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 | 51.53 nDCG@10 x100 (mean of 10 tasks) | 10 tasks, split=test | leaderboard | 2026-10-24 |
| measured MTEB · Embed text for search | 62.07 mean over 40/41 tasks present | 40/41 tasks | leaderboard | 2026-10-24 |
| measured MTEB Reranking · Rerank search results | 44.74 MAP x100 (mean of 2 tasks) | 2 tasks, split=test | leaderboard | 2026-10-24 |
| author-reported Mr. TyDi · Embed text for search | 70.5 | Avg MRR@10 | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 77.5 | ar | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 73.2 | bn | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 60.8 | en | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 66.8 | fi | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 68.5 | id | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 62.5 | ja | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 61.6 | ko | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 65.8 | ru | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 72.7 | sw | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 90.2 | te | model author | 2026-10-07 |
| author-reported Mr. TyDi · Embed text for search | 76.2 | th | model author | 2026-10-07 |
| author-reported ArguAna_default_test (mteb/arguana) | 54.36 | Obtained using MTEB v1.36.26 | Obtained using MTEB v1.36.26 | 2026-10-24 |
| author-reported ArguAna (mteb/arguana) | 54.36 | Obtained using MTEB v1.36.26 | Obtained using MTEB v1.36.26 | 2026-10-24 |
| author-reported BrightAopsRetrieval_default_standard (mteb/BRIGHT) | 7.22 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightAopsRetrieval (mteb/BRIGHT) | 7.22 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightBiologyLongRetrieval_default_long (mteb/BRIGHT) | 1.94 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightBiologyLongRetrieval (mteb/BRIGHT) | 1.94 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightBiologyRetrieval_default_standard (mteb/BRIGHT) | 1.74 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightBiologyRetrieval (mteb/BRIGHT) | 1.74 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEarthScienceLongRetrieval_default_long (mteb/BRIGHT) | 21.55 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEarthScienceLongRetrieval (mteb/BRIGHT) | 21.55 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
Show all 39 rows (14 more)
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported BrightEarthScienceRetrieval_default_standard (mteb/BRIGHT) | 15.06 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEarthScienceRetrieval (mteb/BRIGHT) | 15.06 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEconomicsLongRetrieval_default_long (mteb/BRIGHT) | 13.59 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEconomicsLongRetrieval (mteb/BRIGHT) | 13.59 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEconomicsRetrieval_default_standard (mteb/BRIGHT) | 7.06 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightEconomicsRetrieval (mteb/BRIGHT) | 7.06 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightLeetcodeRetrieval_default_standard (mteb/BRIGHT) | 27.87 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightLeetcodeRetrieval (mteb/BRIGHT) | 27.87 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightPonyLongRetrieval_default_long (mteb/BRIGHT) | 2.34 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightPonyLongRetrieval (mteb/BRIGHT) | 2.34 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightPonyRetrieval_default_standard (mteb/BRIGHT) | 13.02 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightPonyRetrieval (mteb/BRIGHT) | 13.02 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightPsychologyLongRetrieval_default_long (mteb/BRIGHT) | 5.94 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
| author-reported BrightPsychologyLongRetrieval (mteb/BRIGHT) | 5.94 | Obtained using MTEB v2.10.12 | Obtained using MTEB v2.10.12 | 2026-03-31 |
What the card says
1 cited sentence
Summary generated, cited
The model uses 24 layers and an embedding size of 1024 for multilingual tasks.[1]
Generated 2026-10-07 15:00 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.
Use-case evidence
3 use cases
What it is good at 3 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
- multilingual-e5-large has 560M parameters. (safetensors metadata, observed 2026-10-24)
- Context length is 514 tokens (514). (config.json max_position_embeddings, observed 2026-10-24)
- License: MIT (friction: none). Permissive. (license file, observed 2026-10-24)
- Created 2023-06-30, last modified 2026-04-02. (HF API, observed 2026-10-24)
| Pipeline tag | feature-extraction |
|---|---|
| Library | sentence-transformers |
| Architectures | XLMRobertaModel |
| Model type | xlm-roberta |
| Dtype | float32 |
| Params by dtype | F32: 559,890,432 · I64: 514 |
| Size band | S |
| MoE | no |
| Context length | 514 (config.json max_position_embeddings) |
| Languages | multilingual, af, am, ar, as, az, be, bg, bn, br, bs, ca, cs, cy, da, de, el, en, eo, es, et, eu, fa, fi, fr, fy, ga, gd, gl, gu, ha, he, hi, hr, hu, hy, id, is, it, ja, jv, ka, kk, km, kn, ko, ku, ky, la, lo, lt, lv, mg, mk, ml, mn, mr, ms, my, ne, nl, no, om, or, pa, pl, ps, pt, ro, ru, sa, sd, si, sk, sl, so, sq, sr, su, sv, sw, ta, te, th, tl, tr, ug, uk, ur, uz, vi, xh, yi, zh |
| Papers | arXiv:2402.05672, arXiv:2108.08787, arXiv:2104.08663, arXiv:2210.07316 |
| License | MIT (mit) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: supported · XLMRobertaModel · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · XLMRobertaModel · 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 multilingual-e5-large: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 1.3 GB · Edge (<4 GB) |
|---|---|
| FP8 (1 byte/param) | 0.7 GB · Edge (<4 GB) |
| INT4 (0.5 bytes/param) | 0.3 GB · Edge (<4 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
All variants
40 variants
Lineage
finetune: 193quantized: 269adapter: 4
Top variants (40 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| nlpai-lab/KoE5 | finetune | — | 560M | S | mit | nlpai-lab | 52 | 423.2k |
| upskyy/e5-large-korean | finetune | — | 560M | S | mit | upskyy | 3 | 189.7k |
| djovak/embedic-large | finetune | — | 560M | S | mit | djovak | 5 | 116.2k |
| Xenova/multilingual-e5-large | quantized | — | — | — | — | Xenova | 13 | 51.4k |
| mrm8488/multilingual-e5-large-ft-sts-spanish-matryoshka-768-16-5e | finetune | — | 560M | S | — | mrm8488 | 6 | 43.9k |
| aisingapore/SEA-LION-E5-Embedding-600M | finetune | — | 560M | S | mit | aisingapore | 2 | 18.3k |
| Metric-AI/armenian-text-embeddings-2-large | finetune | — | 560M | S | mit | Metric-AI | 7 | 12.2k |
| mjwong/multilingual-e5-large-xnli | finetune | — | 560M | S | mit | mjwong | 6 | 7.2k |
| vgarg/fw_identification_model_e5_large_v5_14_02_24 | finetune | — | 560M | S | — | vgarg | 0 | 5.4k |
| automatedMotiveCoder/setfit | finetune | — | 560M | S | mit | automatedMotiveCoder | 0 | 4.4k |
| louisbrulenaudet/lemone-embed-l | finetune | — | 560M | S | apache-2.0 | louisbrulenaudet | 0 | 4.2k |
| ukr-detect/ukr-emotions-classifier | finetune | — | 560M | S | openrail++ | ukr-detect | 2 | 3.7k |
| phate334/multilingual-e5-large-gguf | quantized | gguf | — | — | mit | phate334 | 2 | 3.6k |
| mjwong/multilingual-e5-large-xnli-anli | finetune | — | 560M | S | cc-by-nc-4.0 | mjwong | 2 | 3.5k |
| mik3ml/multilingual-e5-large-ita | finetune | — | 560M | S | apache-2.0 | mik3ml | 0 | 3.5k |
| andricValdez/multilingual-e5-large-finetuned-autext24 | finetune | — | 560M | S | mit | andricValdez | 0 | 3.4k |
| mainguyen9/vietlegal-e5 | finetune | — | 560M | S | apache-2.0 | mainguyen9 | 2 | 3.1k |
| alphaedge-ai/multilingual-e5-large-nep-16384 | quantized | — | 321M | S | mit | alphaedge-ai | 1 | 2.9k |
| cstr/multilingual-e5-large-GGUF | quantized | GGUF | — | — | mit | cstr | 0 | 2.4k |
| antoinelouis/french-me5-large | quantized | — | 342M | S | mit | antoinelouis | 0 | 2.3k |
| mrm8488/multilingual-e5-large-ft-sts-spanish-matryoshka-768-64-5e | finetune | — | 560M | S | — | mrm8488 | 3 | 2.1k |
| slovak-nlp/e5-sk-large | finetune | — | 365M | S | mit | slovak-nlp | 4 | 2k |
| nnch/multilingual-e5-large-Q4_K_M-GGUF | quantized | GGUF | — | — | mit | nnch | 1 | 2k |
| jester6136/multilingual-e5-large-m2v | finetune | — | 64.0M | S | mit | jester6136 | 0 | 1.7k |
| kietnt0603/nrk-legal-large | finetune | — | 560M | S | apache-2.0 | kietnt0603 | 0 | 1.6k |
| hamtaai/e5-large-hadith-v2 | finetune | — | 560M | S | apache-2.0 | hamtaai | 1 | 1.6k |
| tatsuyaaaaaaa/multilingual-e5-large-gguf | quantized | gguf | — | — | mit | tatsuyaaaaaaa | 0 | 1.3k |
| GiliGold/Knesset-multi-e5-large | finetune | — | 560M | S | — | GiliGold | 0 | 1.3k |
| chris-code/multilingual-e5-large-Q8_0-GGUF | quantized | Q8_0 | — | — | mit | chris-code | 0 | 1.1k |
| julian-schelb/multilingual-e5-large-emb-lat-intertext-v1 | finetune | — | 560M | S | apache-2.0 | julian-schelb | 0 | 1.1k |
| vgarg/fw_identification_model_e5_large_v5_14_12_23 | finetune | — | 560M | S | — | vgarg | 0 | 1k |
| iara-project/e5-large-matryoshka-sts-pt | finetune | — | 560M | S | mit | iara-project | 0 | 965 |
| soichisumi/multilingual-e5-large-Q8_0-GGUF | quantized | Q8_0 | — | — | mit | soichisumi | 1 | 914 |
| SriRamanaAtmic/AtmicEmbeddingv2 | finetune | — | 560M | S | mit | SriRamanaAtmic | 0 | 915 |
| nexus666/multilingual-e5-large-m2v | finetune | — | 128M | S | mit | nexus666 | 1 | 828 |
| vgarg/promo_prescriptive_gpt_28_02_2024_v1 | finetune | — | 560M | S | — | vgarg | 0 | 828 |
| abdulrahman-nuzha/intfloat-multilingual-e5-large-arabic-fp16 | finetune | -fp16 | 560M | S | mit | abdulrahman-nuzha | 0 | 733 |
| DuyAnh1/autotrain-upmei-jyjou | finetune | — | 560M | S | — | DuyAnh1 | 0 | 685 |
| louisbrulenaudet/lemone-embed-l-boost | finetune | — | 560M | S | apache-2.0 | louisbrulenaudet | 0 | 675 |
| Zenabius/multilingual-e5-large-Q8_0-GGUF | quantized | Q8_0 | — | — | mit | Zenabius | 0 | 642 |
Repository changelog
17 commits
Repository changelog 17 commits
| Date | Commit | SHA |
|---|---|---|
| 2026-04-02 | Add evaluation results for model intfloat/multilingual-e5-large revision ab10c1a7f42e74530fe7ae5be82e6d4f11a719eb (#59) | 3d7cfbd |
| 2026-04-02 | Add evaluation results for model intfloat/multilingual-e5-large revision ab10c1a7f42e74530fe7ae5be82e6d4f11a719eb (#58) | aa1c06b |
| 2025-02-17 | Add exported openvino model 'openvino_model.xml' (#51) | 0dc5580 |
| 2025-02-17 | Add exported onnx model 'model_O4.onnx' (#50) | 6560622 |
| 2025-02-17 | Add exported onnx model 'model_qint8_avx512_vnni.onnx' (#49) | 631fbf7 |
| 2024-02-15 | Update README.md | ab10c1a |
| 2024-02-14 | Add link to new paper (#29) | bbc53bb |
| 2023-09-23 | Update README.md (#16) | 9f78368 |
| 2023-08-07 | Update README.md | c505dce |
| 2023-07-27 | add sentence similarity tag | 9228c45 |
| 2023-07-27 | Add 'Sentence Transformers' Tag to generate sentence embedding (#8) | f4abbbe |
| 2023-07-20 | Update README.md | 598f236 |
| 2023-07-14 | Upload 3 files (#5) | 6693d0c |
| 2023-07-03 | Adding ONNX file of this model (#2) | e8061c7 |
| 2023-07-03 | Adding `safetensors` variant of this model (#1) | 21599c7 |
| 2023-06-30 | upload model weights | 9962edc |
| 2023-06-30 | initial commit | b7915d0 |
All providers
2 providers · 2 endpoints
Where to run 2 providers · 2 endpoints
Hosted endpoints observed for multilingual-e5-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 hf, deepinfra | intfloat/multilingual-e5-large | 512 | fp32 | $0.01 | — | — | — | — | live | 2026-10-25 |
| HF Inference hf | intfloat/multilingual-e5-large | — | — | — | — | — | — | — | live | 2026-10-25 |
Articles and coverage
1 piece
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.
Analysis1
emergentmind.com2,514 words
Multilingual-E5-Large uses a 24-layer Transformer initialized from XLM-RoBERTa-large with 560M parameters and employs a shared encoder for queries and documents. The model achieves competitive benchmarks across multilingual tasks, supporting applications like cross-lingual retrieval, bitext mining, and Arabic machine-generated text detection.generated, cited
Papers
4 papers
Papers 4
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).
Similar models
3 models
Similar models
- bge-reranker-large · 560M same size band (S), same task (feature-extraction), fewer providers, permissive license
- Qwen3-Embedding-0.6B · 596M same size band (S), same task (feature-extraction), fewer providers, permissive license
- jina-embeddings-v3 · 572M same size band (S), same task (feature-extraction)
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 multilingual-e5-large
Watch this model
- /feeds/models/intfloat--multilingual-e5-large.xml(RSS 2.0) · every change to multilingual-e5-large
- /feeds/families/intfloat--e5-large.xml(RSS 2.0) · every change across the e5-large family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.