Skip to content

Benchmark

MTEB(Multilingual, v2) - BitextMining

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

Models measured
36
Scores recorded
36
Sources
1
IQ weight
0.52

Score distribution

49.7%
Mean
57.8%
Median
23.4%
Std dev
23.6%
10th percentile
79.4%
90th percentile
0.9%
Min
0.9%91.7%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.52 for embedding models, measured over 36 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Gemini Embedding 2BitextMining· MTEB Leaderboard data91.7%
2Qwen3 Embedding 8BBitextMining· MTEB Leaderboard data80.9%
3Multilingual E5 Large InstructBitextMining· MTEB Leaderboard data80.1%
4Qwen3 Embedding 4BBitextMining· MTEB Leaderboard data79.4%
5Gemini Embedding 001BitextMining· MTEB Leaderboard data79.3%
6Qwen3 Embedding 0.6BBitextMining· MTEB Leaderboard data72.2%
7Text Multilingual Embedding 002BitextMining· MTEB Leaderboard data70.7%
8Jina Embeddings v5 Omni SmallBitextMining· MTEB Leaderboard data69.7%
9Jina Embeddings v5 Text SmallBitextMining· MTEB Leaderboard data69.7%
10E5 Mistral 7B InstructBitextMining· MTEB Leaderboard data68.4%
11Jina Embeddings v5 Omni NanoBitextMining· MTEB Leaderboard data67.7%
12Jina Embeddings v5 Text NanoBitextMining· MTEB Leaderboard data67.7%
13embeddinggemma-2BitextMining· MTEB Leaderboard data65.5%
14Jina Embeddings v3BitextMining· MTEB Leaderboard data65.3%
15snowflake-arctic-embed-l-v2.0BitextMining· MTEB Leaderboard data64.1%
16Text Embedding 3 LargeBitextMining· MTEB Leaderboard data62.1%
17Granite Embedding 278m MultilingualBitextMining· MTEB Leaderboard data58.5%
18NV-Embed-v2BitextMining· MTEB Leaderboard data57.8%
19snowflake-arctic-embed-m-v2.0BitextMining· MTEB Leaderboard data53.7%
20Paraphrase Multilingual MPNet v2BitextMining· MTEB Leaderboard data52.1%
21Text Embedding 3 SmallBitextMining· MTEB Leaderboard data50.4%
22E5 Large V2BitextMining· MTEB Leaderboard data33.8%
23E5 Base V2BitextMining· MTEB Leaderboard data31%
24nomic-embed-text-v1BitextMining· MTEB Leaderboard data28.4%
25nomic-embed-text-v1.5BitextMining· MTEB Leaderboard data27.1%
26UAE-Large-V1BitextMining· MTEB Leaderboard data26.8%
27BGE Large En V1.5BitextMining· MTEB Leaderboard data26.6%
28Jina Embedding B EN v1BitextMining· MTEB Leaderboard data26.3%
29Gte LargeBitextMining· MTEB Leaderboard data26.2%
30Gte BaseBitextMining· MTEB Leaderboard data25.3%
31BGE Base En V1.5BitextMining· MTEB Leaderboard data23.7%
32bge-small-en-v1.5BitextMining· MTEB Leaderboard data23.6%
33All MiniLM L12 V2BitextMining· MTEB Leaderboard data22.9%
34All Mpnet Base V2BitextMining· MTEB Leaderboard data21.2%
35All MiniLM L6 V2BitextMining· MTEB Leaderboard data20.1%
36Jina Embeddings v2 Base ENBitextMining· MTEB Leaderboard data0.9%