MTEB Retrieval

Ranks passages in a corpus by relevance to a query using embedding similarity, scored by nDCG@10 across 15 mostly-English MTEB datasets.

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categoryembedding
Subcategoryretrieval
Page statusactive
MetricnDCG@10
Directionhigher_is_better
Unit%
Dataset size15
Dataset licenceVaries by dataset.
PublisherHugging Face and Cohere (original authors); maintained today by the open embeddings-benchmark community

What it measures

Given a short query, the model embeds the query and every candidate passage in a fixed corpus, and passages are ranked by similarity with no task-specific fine-tuning. It is the MTEB task type closest to production semantic search and retrieval-augmented generation.

Task format

Query-to-corpus ranking: embed a query and a candidate pool, rank the pool by cosine or dot-product similarity, score against relevance judgements.

Models reporting this benchmark

These figures come from the model cards, which carry one collection date per card and no per-score attribution. They are shown as reported, not as verified evidence.
ModelProviderScoreCard as of
bge reranker v2 m3BAAI68.52026-04
NV Embed v2NVIDIA67.22026-04
jina reranker v2 base multilingualJina AI66.22026-04
bge m3BAAI65.82026-04
Qwen3 Embedding 8BAlibaba / Qwen Team65.52026-04
Qwen3 VL Embedding 8BAlibaba / Qwen Team65.52026-04
jina embeddings v4Jina AI65.02026-04
jina embeddings v4 vllm retrievalJina AI65.02026-04
Voyage 3Voyage AI64.52026-04
jina embeddings v3Jina AI63.82026-04
bge multilingual gemma2BAAI63.52026-04
Gemini Embedding 001Google DeepMind63.52026-04
e5 mistral 7B instructintfloat63.22026-04
snowflake arctic embed l v2.0Snowflake63.02026-04
Qwen3 Embedding 4BAlibaba / Qwen Team62.52026-04
SFR Embedding 2 RSalesforce62.52026-04
nomic embed text v2 moeNomic AI62.22026-04
nomic embed text v2 moe GGUFNomic AI62.22026-04
text-embedding-3-largeOpenAI62.12026-04
multilingual e5 large instructintfloat61.22026-04
bge large zh v1.5BAAI612026-04
bge large en v1.5BAAI60.52026-04
snowflake arctic embed lSnowflake60.52026-04
granite embedding english r2IBM60.22026-04
granite embedding small english r2IBM60.22026-04
bge large zhBAAI602026-04
Voyage Code 3Voyage AI60.02026-04
Voyage Finance 2Voyage AI59.82026-04
snowflake arctic embed m v2.0Snowflake59.52026-04
Voyage Multilingual 2Voyage AI59.52026-04
bge large enBAAI59.22026-04
Voyage Law 2Voyage AI59.22026-04
nomic embed text v1.5Nomic AI59.12026-04
nomic embed text v1.5 GGUFNomic AI59.12026-04
e5 large v2intfloat58.82026-04
Voyage 3 LiteVoyage AI58.22026-04
snowflake arctic embed m v1.5Snowflake58.02026-04
bge base en v1.5BAAI57.82026-04
granite embedding 278M multilingualIBM57.82026-04
nomic embed text v1Nomic AI57.82026-04
bge base zh v1.5BAAI57.52026-04
multilingual e5 largeintfloat57.52026-04
snowflake arctic embed m longSnowflake57.52026-04
bge code v1BAAI57.22026-04
e5 largeintfloat57.22026-04
nomic embed codeNomic AI57.02026-04
bge base zhBAAI56.82026-04
snowflake arctic embed mSnowflake56.82026-04
bge base enBAAI56.52026-04
e5 large unsupervisedintfloat56.52026-04
jina embeddings v2 base enJina AI56.52026-04
Qwen3 Embedding 0.6BAlibaba / Qwen Team56.52026-04
Qwen3 VL Embedding 2BAlibaba / Qwen Team56.52026-04
jina embeddings v2 base deJina AI562026-04
Mistral EmbedMistral AI55.82026-04
SFR Embedding Code 400M RSalesforce55.82026-04
e5 base v2intfloat55.52026-04
jina embeddings v2 base codeJina AI55.52026-04
multilingual e5 baseintfloat55.02026-04
granite embedding 125M englishIBM54.52026-04
bge small en v1.5BAAI54.22026-04
e5 baseintfloat54.22026-04
bge small zh v1.5BAAI542026-04
granite embedding 107M multilingualIBM54.02026-04
jina embeddings v2 small enJina AI542026-04
snowflake arctic embed sSnowflake53.52026-04
e5 small v2intfloat53.22026-04
text-embedding-3-smallOpenAI53.22026-04
bge small zhBAAI532026-04
multilingual e5 smallintfloat53.02026-04
bge small enBAAI52.02026-04
e5 smallintfloat51.82026-04
snowflake arctic embed xsSnowflake50.52026-04
all mpnet base v2Sentence Transformers50.22026-04
text-embedding-ada-002OpenAI49.82026-04
all MiniLM L12 v2Sentence Transformers49.22026-04
all MiniLM L6 v2Sentence Transformers48.52026-04
granite embedding 30M englishIBM48.52026-04
granite embedding 30M sparseIBM48.52026-04
multi qa mpnet base dot v1Sentence Transformers48.52026-04
multi qa mpnet base cos v1Sentence Transformers48.22026-04
all roberta large v1Sentence Transformers48.02026-04
all distilroberta v1Sentence Transformers47.52026-04
paraphrase mpnet base v2Sentence Transformers47.02026-04
multi qa MiniLM L6 cos v1Sentence Transformers46.82026-04
msmarco bert base dot v5Sentence Transformers46.52026-04
paraphrase multilingual mpnet base v2Sentence Transformers46.22026-04
msmarco MiniLM L12 cos v5Sentence Transformers45.52026-04
paraphrase multilingual MiniLM L12 v2Sentence Transformers44.82026-04
paraphrase MiniLM L12 v2Sentence Transformers44.52026-04
msmarco MiniLM L6 v3Sentence Transformers44.22026-04
paraphrase MiniLM L6 v2Sentence Transformers44.02026-04
LaBSESentence Transformers43.52026-04
paraphrase MiniLM L3 v2Sentence Transformers41.52026-04
distiluse base multilingual cased v2Sentence Transformers40.52026-04
distiluse base multilingual cased v1Sentence Transformers38.82026-04

Data

This page as JSON · Edit on GitHub