Skip to content

Benchmark

MTEB(Law, v1) - Retrieval

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

Models measured
32
Scores recorded
32
Sources
1
IQ weight
0.99

Score distribution

47.6%
Mean
44.3%
Median
13.9%
Std dev
35.8%
10th percentile
64.3%
90th percentile
3.2%
Min
3.2%71.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.99 for embedding models, measured over 32 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 8BRetrieval· MTEB Leaderboard data71.7%
2Qwen3 Embedding 4BRetrieval· MTEB Leaderboard data69.6%
3Nemotron 3 Embed 1bRetrieval· MTEB Leaderboard data68.6%
4Jina Embeddings v5 Text SmallRetrieval· MTEB Leaderboard data64.3%
5Qwen3 Embedding 0.6BRetrieval· MTEB Leaderboard data64.2%
6INF Retriever v1Retrieval· MTEB Leaderboard data63.7%
7Text Embedding 3 LargeRetrieval· MTEB Leaderboard data59.4%
8snowflake-arctic-embed-l-v2.0Retrieval· MTEB Leaderboard data59%
9Jina Embeddings v3Retrieval· MTEB Leaderboard data56.9%
10snowflake-arctic-embed-m-v2.0Retrieval· MTEB Leaderboard data56.6%
11Text Embedding 3 SmallRetrieval· MTEB Leaderboard data56.3%
12Multilingual E5 Large InstructRetrieval· MTEB Leaderboard data54.9%
13E5 Mistral 7B InstructRetrieval· MTEB Leaderboard data54.3%
14Multilingual E5 LargeRetrieval· MTEB Leaderboard data48.4%
15Granite Embedding 278m MultilingualRetrieval· MTEB Leaderboard data45.8%
16NV-Embed-v2Retrieval· MTEB Leaderboard data44.3%
17nomic-embed-text-v1Retrieval· MTEB Leaderboard data41.8%
18E5 Large V2Retrieval· MTEB Leaderboard data41.7%
19Gte LargeRetrieval· MTEB Leaderboard data41.2%
20Jina Embedding B EN v1Retrieval· MTEB Leaderboard data40.8%
21BGE Base En V1.5Retrieval· MTEB Leaderboard data40.2%
22Gte BaseRetrieval· MTEB Leaderboard data40%
23BGE Large En V1.5Retrieval· MTEB Leaderboard data39.6%
24E5 Base V2Retrieval· MTEB Leaderboard data39.5%
25UAE-Large-V1Retrieval· MTEB Leaderboard data39.3%
26bge-small-en-v1.5Retrieval· MTEB Leaderboard data38.2%
27All Mpnet Base V2Retrieval· MTEB Leaderboard data37.4%
28Paraphrase Multilingual MPNet v2Retrieval· MTEB Leaderboard data36.4%
29nomic-embed-text-v1.5Retrieval· MTEB Leaderboard data35.8%
30All MiniLM L6 V2Retrieval· MTEB Leaderboard data35.4%
31All MiniLM L12 V2Retrieval· MTEB Leaderboard data33.9%
32Jina Embeddings v2 Base ENRetrieval· MTEB Leaderboard data3.2%