MTEB Classification

Trains a logistic regression probe on a model's embeddings and scores accuracy on 12 classification datasets in English and other languages.

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
Subcategoryclassification
Page statusactive
Metricaccuracy
Directionhigher_is_better
Unit%
Dataset size12
Dataset licenceVaries by dataset.
PublisherHugging Face and Cohere (original authors); maintained today by the open embeddings-benchmark community

What it measures

A labelled train split is embedded, a logistic regression classifier is fit on those embeddings, and accuracy is measured on the embedded test split. The embedding model itself is never fine-tuned; only the small linear probe on top of it is trained.

Task format

Embed train and test splits; fit a scikit-learn logistic regression classifier (max_iter=100) on train embeddings; score accuracy on test embeddings.

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
NV Embed v2NVIDIA73.52026-04
Qwen3 Embedding 8BAlibaba / Qwen Team72.52026-04
Qwen3 VL Embedding 8BAlibaba / Qwen Team72.52026-04
bge m3BAAI72.12026-04
jina embeddings v4Jina AI72.02026-04
jina embeddings v4 vllm retrievalJina AI72.02026-04
Voyage 3Voyage AI71.82026-04
e5 mistral 7B instructintfloat71.52026-04
bge multilingual gemma2BAAI71.22026-04
jina embeddings v3Jina AI71.22026-04
snowflake arctic embed l v2.0Snowflake71.02026-04
bge large zh v1.5BAAI70.82026-04
Gemini Embedding 001Google DeepMind70.82026-04
SFR Embedding 2 RSalesforce70.82026-04
nomic embed text v2 moeNomic AI70.52026-04
nomic embed text v2 moe GGUFNomic AI70.52026-04
Qwen3 Embedding 4BAlibaba / Qwen Team70.52026-04
text-embedding-3-largeOpenAI70.52026-04
bge large en v1.5BAAI70.32026-04
bge large zhBAAI69.82026-04
multilingual e5 large instructintfloat69.82026-04
granite embedding english r2IBM69.52026-04
granite embedding small english r2IBM69.52026-04
bge large enBAAI69.12026-04
snowflake arctic embed lSnowflake69.12026-04
nomic embed text v1.5Nomic AI68.82026-04
nomic embed text v1.5 GGUFNomic AI68.82026-04
Voyage Finance 2Voyage AI68.82026-04
bge base en v1.5BAAI68.52026-04
snowflake arctic embed m v2.0Snowflake68.52026-04
Voyage Code 3Voyage AI68.52026-04
bge base zh v1.5BAAI68.22026-04
e5 large v2intfloat68.22026-04
Voyage Multilingual 2Voyage AI68.22026-04
Voyage Law 2Voyage AI68.02026-04
Voyage 3 LiteVoyage AI67.82026-04
bge base zhBAAI67.52026-04
granite embedding 278M multilingualIBM67.52026-04
nomic embed text v1Nomic AI67.52026-04
multilingual e5 largeintfloat67.32026-04
bge base enBAAI67.22026-04
snowflake arctic embed m v1.5Snowflake67.22026-04
e5 largeintfloat67.02026-04
snowflake arctic embed m longSnowflake67.02026-04
bge code v1BAAI66.82026-04
snowflake arctic embed mSnowflake66.52026-04
e5 large unsupervisedintfloat66.22026-04
jina embeddings v2 base enJina AI66.22026-04
nomic embed codeNomic AI66.22026-04
Mistral EmbedMistral AI66.12026-04
Qwen3 Embedding 0.6BAlibaba / Qwen Team66.02026-04
Qwen3 VL Embedding 2BAlibaba / Qwen Team66.02026-04
bge small en v1.5BAAI65.82026-04
e5 base v2intfloat65.82026-04
jina embeddings v2 base deJina AI65.82026-04
bge small zh v1.5BAAI65.52026-04
multilingual e5 baseintfloat65.52026-04
SFR Embedding Code 400M RSalesforce65.22026-04
jina embeddings v2 base codeJina AI652026-04
bge small zhBAAI64.82026-04
e5 baseintfloat64.82026-04
text-embedding-3-smallOpenAI64.82026-04
all mpnet base v2Sentence Transformers64.52026-04
granite embedding 125M englishIBM64.52026-04
jina embeddings v2 small enJina AI64.22026-04
bge small enBAAI64.02026-04
granite embedding 107M multilingualIBM64.02026-04
snowflake arctic embed sSnowflake64.02026-04
all MiniLM L12 v2Sentence Transformers63.82026-04
e5 small v2intfloat63.82026-04
multilingual e5 smallintfloat63.52026-04
all MiniLM L6 v2Sentence Transformers63.22026-04
all roberta large v1Sentence Transformers62.82026-04
e5 smallintfloat62.52026-04
multi qa mpnet base dot v1Sentence Transformers62.52026-04
multi qa mpnet base cos v1Sentence Transformers62.22026-04
paraphrase mpnet base v2Sentence Transformers62.22026-04
text-embedding-ada-002OpenAI62.12026-04
all distilroberta v1Sentence Transformers62.02026-04
snowflake arctic embed xsSnowflake62.02026-04
paraphrase multilingual mpnet base v2Sentence Transformers61.52026-04
multi qa MiniLM L6 cos v1Sentence Transformers61.02026-04
paraphrase multilingual MiniLM L12 v2Sentence Transformers60.52026-04
paraphrase MiniLM L12 v2Sentence Transformers60.22026-04
granite embedding 30M englishIBM60.02026-04
granite embedding 30M sparseIBM60.02026-04
paraphrase MiniLM L6 v2Sentence Transformers59.82026-04
msmarco bert base dot v5Sentence Transformers59.52026-04
LaBSESentence Transformers59.22026-04
msmarco MiniLM L12 cos v5Sentence Transformers59.02026-04
msmarco MiniLM L6 v3Sentence Transformers58.02026-04
paraphrase MiniLM L3 v2Sentence Transformers57.52026-04
distiluse base multilingual cased v2Sentence Transformers57.02026-04
distiluse base multilingual cased v1Sentence Transformers55.52026-04

Data

This page as JSON · Edit on GitHub