Skip to content

Benchmark

MTEB(Multilingual, v1) - Clustering

Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.

Models measured
31
Scores recorded
31
Sources
1
IQ weight
1.58

Score distribution

45.4%
Mean
42.6%
Median
7.2%
Std dev
40.5%
10th percentile
54.1%
90th percentile
22.1%
Min
22.1%58.1%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 1.58 for embedding models, measured over 31 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Qwen3 Embedding 8BClustering· MTEB Leaderboard data58.1%
2Qwen3 Embedding 4BClustering· MTEB Leaderboard data57.7%
3Gemini Embedding 001Clustering· MTEB Leaderboard data55%
4Jina Embeddings v5 Omni SmallClustering· MTEB Leaderboard data54.1%
5Jina Embeddings v5 Text SmallClustering· MTEB Leaderboard data54.1%
6Jina Embeddings v5 Omni NanoClustering· MTEB Leaderboard data53.4%
7Qwen3 Embedding 0.6BClustering· MTEB Leaderboard data52.9%
8Multilingual E5 Large InstructClustering· MTEB Leaderboard data51.5%
9E5 Mistral 7B InstructClustering· MTEB Leaderboard data51.4%
10Text Multilingual Embedding 002Clustering· MTEB Leaderboard data48.5%
11Text Embedding 3 LargeClustering· MTEB Leaderboard data47.6%
12Text Embedding 3 SmallClustering· MTEB Leaderboard data46.8%
13Jina Embeddings v3Clustering· MTEB Leaderboard data46.4%
14snowflake-arctic-embed-l-v2.0Clustering· MTEB Leaderboard data43.6%
15snowflake-arctic-embed-m-v2.0Clustering· MTEB Leaderboard data43%
16UAE-Large-V1Clustering· MTEB Leaderboard data42.6%
17Granite Embedding 278m MultilingualClustering· MTEB Leaderboard data42.3%
18BGE Large En V1.5Clustering· MTEB Leaderboard data42.2%
19Gte LargeClustering· MTEB Leaderboard data42.1%
20Gte BaseClustering· MTEB Leaderboard data42.1%
21BGE Base En V1.5Clustering· MTEB Leaderboard data42.1%
22BGE M3Clustering· MTEB Leaderboard data41.8%
23nomic-embed-text-v1Clustering· MTEB Leaderboard data41.6%
24NV-Embed-v2Clustering· MTEB Leaderboard data41.4%
25Paraphrase Multilingual MPNet v2Clustering· MTEB Leaderboard data41.1%
26E5 Base V2Clustering· MTEB Leaderboard data40.8%
27E5 Large V2Clustering· MTEB Leaderboard data40.6%
28bge-small-en-v1.5Clustering· MTEB Leaderboard data40.5%
29nomic-embed-text-v1.5Clustering· MTEB Leaderboard data40.5%
30All MiniLM L6 V2Clustering· MTEB Leaderboard data38%
31Jina Embeddings v2 Base ENClustering· MTEB Leaderboard data22.1%