Benchmark · Embeddings

MTEB English v2

An embedding model's mean main score across the 41 English tasks of MTEB(eng, v2), covering retrieval, classification, clustering, reranking, pair classification, semantic similarity and summarisation, as the MTEB leaderboard's Mean (Task) column.

Score
mean task score (0 to 100), higher is better
Run by
MTEB maintainers (an open community project started at Hugging Face), from results submitted by pull request
Status
Current, newest result 20 November 2026
Updated
continuous (pull requests to the results repository)
Data
the published file, CC0-1.0, read 20 November 2026
Method
huggingface.co/spaces/mteb/leaderboard

Results

11 models we list, each at its best setting, best first.

#ModelScoreRangeWho ran itSettingDate
1Qwen3-Embedding-8B BBSubmitted by the makersource
2Qwen3-Embedding-4B BBSubmitted by the makersource
3Gemini Embedding BSubmitted by the makermodel gemini-embedding-001source
4jina-embeddings-v5-omni-small DSubmitted by the makersource
5Jina Embeddings and Reranker BSubmitted by the makermodel jina-embeddings-v5-omni-nanosource
6Qwen3-Embedding-0.6B BSubmitted by the makersource
7EmbeddingGemma 300M DSubmitted by the makersource
8EmbeddingGemma 2 BSubmitted by the makersource
9Cohere Embed and Rerank BBSubmitted by the makermodel Cohere-embed-english-v3.0source
10Nomic Embed ESubmitted by the makermodel nomic-embed-text-v1source
11Granite Embedding Small English R2 ESubmitted by the makersource
Every setting, 16 rows
ModelAs the leaderboard names itScoreWho ran itSettingDate
Qwen3-Embedding-8BQwen/Qwen3-Embedding-8BSubmitted by the makersource
Qwen3-Embedding-4BQwen/Qwen3-Embedding-4BSubmitted by the makersource
Gemini Embeddinggoogle/gemini-embedding-001Submitted by the makermodel gemini-embedding-001source
jina-embeddings-v5-omni-smalljinaai/jina-embeddings-v5-omni-smallSubmitted by the makersource
Jina Embeddings and Rerankerjinaai/jina-embeddings-v5-omni-nanoSubmitted by the makermodel jina-embeddings-v5-omni-nanosource
Jina Embeddings and Rerankerjinaai/jina-embeddings-v5-text-nanoSubmitted by the makermodel jina-embeddings-v5-text-nanosource
Qwen3-Embedding-0.6BQwen/Qwen3-Embedding-0.6BSubmitted by the makersource
EmbeddingGemma 300Mgoogle/embeddinggemma-300mSubmitted by the makersource
EmbeddingGemma 2google/embeddinggemma-2Submitted by the makersource
Cohere Embed and RerankCohere/Cohere-embed-english-v3.0Submitted by the makermodel Cohere-embed-english-v3.0source
Cohere Embed and RerankCohere/Cohere-embed-multilingual-v3.0Submitted by the makermodel Cohere-embed-multilingual-v3.0source
Cohere Embed and RerankCohere/Cohere-embed-english-light-v3.0Submitted by the makermodel Cohere-embed-english-light-v3.0source
Nomic Embednomic-ai/nomic-embed-text-v1Submitted by the makermodel nomic-embed-text-v1source
Cohere Embed and RerankCohere/Cohere-embed-multilingual-light-v3.0Submitted by the makermodel Cohere-embed-multilingual-light-v3.0source
Nomic Embednomic-ai/nomic-embed-text-v1.5Submitted by the makermodel nomic-embed-text-v1.5source
Granite Embedding Small English R2ibm-granite/granite-embedding-small-english-r2Submitted by the makersource

A rank is among the models we list with a result, not a place on the whole leaderboard, and a model offered by several hosts is shown once, on its own listing. These results don't move our grades. Data: MTEB, embeddings-benchmark/results (CC0), CC0-1.0.

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.