s

All MiniLM L12 v2

Aug 2021
sAll MiniLM L12 v2
Compare in
on
$0.005per 1M tokens
43.7MTEB English v2
Aug 2021released
384Dimensions
0K tokensMax input
enLanguages

Rankingconfiguration unstated?

Full ranking →
1F2LLM v2 8B59.32F2LLM v2 14B59.13Text Embedding 3 Large56.84F2LLM v2 4B56.75BOOM 4B v155.4
···
24E5 Large v250.225All MiniLM L12 v249.726Multilingual E5 Base49.5

Strong at

C
Embeddings & Retrieval#132 of 200

Alternatives?

ModelScoreCostPriceWhat changes
BetterText Embedding 3 Smallunstated+6%+300%$0.02Languages: not verifiedcOpenF2LLM v2 8Bunstated+19%——Memory (4-bit): 0 GB → 4.5 GBBetterAmazon Titan Embed Text v2+5%+300%$0.02Languages: not verifiedBetterCohere Embed English v3+8%+1,900%$0.10Languages: not verifiedBetterText Embedding Ada 002unstated+5%+1,900%$0.10Languages: not verified

Run it yourself

estimate
Fits on✓ 8 GB (phone, laptop) · full quality✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use

Benchmarks

in Embeddings & Retrieval
BEIR · Mean over tasks · English42.7#139
BEIR · Mean over task types · English42.7#139
BEIR · Retrieval · English42.7#139
MTEB Mandarin Chinese v1 · Retrieval · Mandarin Chinese6#84
MTEB Mandarin Chinese v1 · Reranking · Mandarin Chinese26.3#76
MTEB Mandarin Chinese v1 · Classification · Mandarin Chinese43#42
ChemTEB · Mean over tasks · multilingual (13 languages)69#32
ChemTEB · Mean over task types · multilingual (13 languages)49.7#25
ChemTEB · Classification · English80.2#32
ChemTEB · Retrieval · English64.1#40
ChemTEB · Pair Classification · multilingual (13 languages)54.4#37
ChemTEB · Clustering · English49.7#11
ChemTEB · Bitext Mining · English0.1#37
JMTEB v2 · STS · Japanese66.9#78
LongEmbed · Mean over tasks · English25.3#142
LongEmbed · Mean over task types · English25.3#142
LongEmbed · Retrieval · English25.3#142
MTEB German v1 · Mean over tasks · German37.1#71
MTEB German v1 · Mean over task types · German36#71
MTEB German v1 · Clustering · German18.4#115
MTEB German v1 · Retrieval · German32#78
MTEB German v1 · Classification · German49.1#100
MTEB German v1 · STS · German44.5#105
MTEB German v1 · Pair Classification · German48.4#151
MTEB German v1 · Reranking · German23.3#91
MTEB Europe v1 · STS · multilingual (8 languages)64.2#54
MTEB Europe v1 · Bitext Mining · multilingual (28 languages)32.1#140
MTEB Europe v1 · Instruction Reranking · English-0.8#81
MTEB Europe v1 · Pair Classification · multilingual (11 languages)81.5#101
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)7.6#144
MTEB French v1 · Retrieval · multilingual (2 languages)35.9#77
MTEB French v1 · Reranking · French68.1#71
MTEB French v1 · Clustering · French34#93
MTEB French v1 · Summarization · French26.6#148
MTEB Indic v1 · Retrieval · multilingual (14 languages)13.9#100
MTEB Indic v1 · Bitext Mining · multilingual (20 languages)3.5#112
MTEB Indic v1 · Pair Classification · Hindi55#114
MTEB Indic v1 · STS · multilingual (13 languages)-5.3#194
MTEB Korean v2 · STS · Korean38.6#63
MTEB Korean v2 · Reranking · Korean8.8#114
MTEB Korean v2 · Classification · Korean21.7#101
MTEB Dutch v1 · Mean over tasks · Dutch36.7#96
MTEB Dutch v1 · Mean over task types · Dutch45.2#96
MTEB Dutch v1 · Classification · Dutch45.5#98
MTEB Dutch v1 · Clustering · Dutch18.2#104
MTEB Dutch v1 · Retrieval · Dutch31.9#95
MTEB Dutch v1 · STS · Dutch59.8#86
MTEB Dutch v1 · Multilabel Classification · Dutch27.1#106
MTEB Dutch v1 · Pair Classification · Dutch58.4#121
MTEB Dutch v1 · Reranking · Dutch75.2#154
MTEB Polish v1 · Pair Classification · Polish64.6#97
MTEB Portuguese v1 · Classification · Portuguese48.4#25
MTEB Portuguese v1 · Multilabel Classification · Portuguese19.3#103
MTEB Law v1 · Mean over tasks · multilingual (3 languages)33.9#127
MTEB Law v1 · Mean over task types · multilingual (3 languages)33.9#127
MTEB Law v1 · Retrieval · multilingual (3 languages)33.9#127
MTEB Medical v1 · Mean over tasks · multilingual (10 languages)33.2#90
MTEB Medical v1 · Mean over task types · multilingual (10 languages)30.2#90
MTEB Medical v1 · Retrieval · multilingual (10 languages)34#92
MTEB Medical v1 · Clustering · English36.6#81
MTEB Medical v1 · Reranking · Mandarin Chinese20.1#137
FollowIR · Mean over tasks · English-0.8#81
FollowIR · Mean over task types · English-0.8#81
FollowIR · Instruction Reranking · English-0.8#81
MTEB Spanish v1 · Clustering · Spanish32.9#118
MTEB Spanish v1 · Reranking · Spanish28.1#101
MTEB Thai v1 · Pair Classification · Thai51.7#132
MTEB Scandinavian v1 · Clustering · multilingual (2 languages)24.7#141
MTEB Scandinavian v1 · Retrieval · multilingual (3 languages)32#49
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)63.5#104
VN-MTEB Vietnamese v1 · Mean over tasks · Vietnamese30.8#25
VN-MTEB Vietnamese v1 · Mean over task types · Vietnamese44.3#25
VN-MTEB Vietnamese v1 · Classification · Vietnamese44.2#29
VN-MTEB Vietnamese v1 · STS · Vietnamese62.3#30
VN-MTEB Vietnamese v1 · Retrieval · Vietnamese14.1#27
VN-MTEB Vietnamese v1 · Clustering · Vietnamese24.4#29
VN-MTEB Vietnamese v1 · Reranking · Vietnamese52.3#29
VN-MTEB Vietnamese v1 · Pair Classification · Vietnamese68.5#26
MTEB Russian v1.1 · Reranking · Russian28.5#93
MTEB Russian v1.1 · Clustering · Russian13.8#139
MTEB Russian v1.1 · Pair Classification · Russian46.4#174
MTEB Russian v1.1 · Multilabel Classification · Russian26#121
RTEB Finnish beta · Mean over tasks · English43.3#77
RTEB French beta · Mean over tasks · multilingual (2 languages)12.9#157
RTEB German beta · Mean over tasks · German7.4#190
RTEB Health beta · Mean over tasks · multilingual (4 languages)35.8#73
RTEB Law beta · Mean over tasks · multilingual (4 languages)25.6#174
MTEB multilingual v2 · STS · multilingual (35 languages)57.2#88
MTEB multilingual v2 · Bitext Mining · multilingual (1026 languages)22.9#133
MTEB multilingual v2 · Instruction Reranking · English-0.8#81
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)14.6#145

FAQ

How much does All MiniLM L12 v2 cost?›

$0.005 per 1M tokens.

Is All MiniLM L12 v2 free to use?›

There is no free API; the cheapest paid price is $0.005 per 1M input tokens. The weights are open, so you can run it yourself.

How good is All MiniLM L12 v2, and what is it best at?›

#132 of 200 in Embeddings.

Is All MiniLM L12 v2 better than E5 Large v2?›

On ChemTEB · Mean over task types · multilingual (13 languages), All MiniLM L12 v2 scores 49.7 vs 50.2 — about the same. It is 50% cheaper.

How can I use All MiniLM L12 v2?›

Through the sentence-transformers API.

When was All MiniLM L12 v2 released? Is there a newer version?›

Released 2021-08-30. It is the current version in its line.

Is All MiniLM L12 v2 open source?›

Yes, the weights are open (Apache-2.0 licence).

Can I run All MiniLM L12 v2 locally?›

The weights are open and can be self-hosted; hardware needs are not confirmed yet.

What is All MiniLM L12 v2?›

All MiniLM L12 v2 is an open-weights embeddings & retrieval model from sentence-transformers, released Aug 2021.

What are the best alternatives to All MiniLM L12 v2?›

Text Embedding 3 Small (better), F2LLM v2 8B (open weights).