Text Embedding 3 Small

Jan 2024
Text Embedding 3 Small
Compare in
on
$0.02per 1M tokens
53.5MTEB English v2
Jan 2024released
1,536Dimensions
8K tokensMax input

Rankingconfiguration unstated?

Full ranking →
1Multilingual E5 Large Instruct80.42F2LLM v2 14B80.33F2LLM v2 8B79.84Solon Embeddings Large 0.1795F2LLM v2 4B78.1
···
61Sgpt Bloom 7b1 Msmarco12.762Text Embedding 3 Small12.263Udever Bloom 1b111.1

Strong at

B
Embeddings & Retrieval#74 of 200

Alternatives?

ModelScoreCostPriceWhat changes
iCheaperMultilingual E5 Large InstructunstatedOpen×6.6-50%$0.01Max input: 8K tokens → 1K tokensiCheaperMultilingual E5 LargeunstatedOpen×6.4-50%$0.01Max input: 8K tokens → 1K tokensiCheaperE5 Large v2unstatedOpen-38%-50%$0.01Max input: 8K tokens → 1K tokensiCheaperE5 Base v2unstatedOpen-46%-75%$0.005Max input: 8K tokens → 1K tokenssCheaperAll Mpnet Base v2unstatedOpen-69%-75%$0.005Max input: 8K tokens → 0K tokens

Benchmarks

in Embeddings & Retrieval
BEIR · Mean over tasks · English51.1#93
BEIR · Mean over task types · English51.1#93
BEIR · Retrieval · English51.1#93
ChemTEB · Mean over tasks · multilingual (13 languages)73.2#7
ChemTEB · Mean over task types · multilingual (13 languages)52.5#12
ChemTEB · Classification · English84.5#10
ChemTEB · Retrieval · English67.9#35
ChemTEB · Pair Classification · multilingual (13 languages)61.1#11
ChemTEB · Clustering · English48.9#15
ChemTEB · Bitext Mining · English0.2#16
MTEB Persian v2 · Mean over tasks · Persian54.9#25
MTEB Persian v2 · Mean over task types · Persian59#25
MTEB Persian v2 · STS · Persian69.1#33
MTEB Persian v2 · Clustering · Persian57.5#30
MTEB Persian v2 · Classification · Persian61.7#28
MTEB Persian v2 · Retrieval · Persian36.6#25
MTEB Persian v2 · Reranking · Persian57.9#51
MTEB Persian v2 · Bitext Mining · Persian51.5#26
MTEB Persian v2 · Pair Classification · Persian78.4#30
JMTEB v2 · Clustering · Japanese50.1#34
JMTEB v2 · STS · Japanese79.5#48
LongEmbed · Mean over tasks · English53.6#43
LongEmbed · Mean over task types · English53.6#43
LongEmbed · Retrieval · English53.6#43
MTEB German v1 · Pair Classification · German54.6#27
MTEB Europe v1 · Bitext Mining · multilingual (28 languages)83.7#64
MTEB Europe v1 · Clustering · multilingual (29 languages)44.1#44
MTEB Europe v1 · Instruction Reranking · English-2.4#141
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)15.2#61
MTEB French v1 · Reranking · French80.1#26
MTEB French v1 · Pair Classification · French61.5#56
MTEB Indic v1 · Retrieval · multilingual (14 languages)55.4#57
MTEB Indic v1 · Bitext Mining · multilingual (20 languages)12.2#62
MTEB Indic v1 · Pair Classification · Hindi62.5#86
MTEB Indic v1 · STS · multilingual (13 languages)4.1#134
MTEB Indic v1 · Clustering · multilingual (21 languages)7.2#105
MTEB Polish v1 · Pair Classification · Polish77.7#58
MTEB Polish v1 · Classification · Polish57.6#40
MTEB Portuguese v1 · Multilabel Classification · Portuguese22.1#38
MTEB Law v1 · Mean over tasks · multilingual (3 languages)56.3#29
MTEB Law v1 · Mean over task types · multilingual (3 languages)56.3#29
MTEB Law v1 · Retrieval · multilingual (3 languages)56.3#29
MTEB Medical v1 · Mean over tasks · multilingual (10 languages)55.6#30
MTEB Medical v1 · Mean over task types · multilingual (10 languages)53.1#34
MTEB Medical v1 · Retrieval · multilingual (10 languages)59#30
MTEB Medical v1 · Clustering · English35.6#114
MTEB Medical v1 · Reranking · Mandarin Chinese64.5#86
FollowIR · Mean over tasks · English-2.4#141
FollowIR · Mean over task types · English-2.4#141
FollowIR · Instruction Reranking · English-2.4#141
MTEB Spanish v1 · Mean over tasks · multilingual (2 languages)65.8#13
MTEB Spanish v1 · Mean over task types · multilingual (2 languages)66.1#13
MTEB Spanish v1 · Pair Classification · Spanish65.8#60
MTEB Spanish v1 · STS · multilingual (2 languages)79.1#14
MTEB Spanish v1 · Retrieval · multilingual (2 languages)53.5#11
MTEB Spanish v1 · Clustering · Spanish49.1#16
MTEB Spanish v1 · Reranking · Spanish59.3#35
MTEB Spanish v1 · Classification · Spanish89.6#12
MTEB Thai v1 · Clustering · Thai16.7#108
MTEB Thai v1 · Pair Classification · Thai61.2#97
MTEB Scandinavian v1 · Mean over tasks · multilingual (6 languages)58.4#20
MTEB Scandinavian v1 · Mean over task types · multilingual (6 languages)58.7#22
MTEB Scandinavian v1 · Clustering · multilingual (2 languages)40.8#46
MTEB Scandinavian v1 · Classification · multilingual (6 languages)60.8#34
MTEB Scandinavian v1 · Retrieval · multilingual (3 languages)66.5#20
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)66.9#72
MTEB Russian v1.1 · Classification · Russian58.3#53
MTEB Russian v1.1 · Clustering · Russian52.7#54
MTEB Russian v1.1 · Pair Classification · Russian53.7#100
MTEB Russian v1.1 · Multilabel Classification · Russian33.3#54
RTEB Finnish beta · Mean over tasks · English60.8#40
RTEB French beta · Mean over tasks · multilingual (2 languages)44.2#59
RTEB German beta · Mean over tasks · German54.8#53
RTEB Health beta · Mean over tasks · multilingual (4 languages)52.3#46
RTEB Law beta · Mean over tasks · multilingual (4 languages)46.8#55
MTEB multilingual v2 · Clustering · multilingual (208 languages)46.1#49
MTEB multilingual v2 · Bitext Mining · multilingual (1026 languages)50.4#82
MTEB multilingual v2 · Instruction Reranking · English-2.4#141
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)19.9#63

FAQ

How much does Text Embedding 3 Small cost?›

$0.02 per 1M tokens.

Is Text Embedding 3 Small free to use?›

There is no free API; the cheapest paid price is $0.02 per 1M input tokens.

How good is Text Embedding 3 Small, and what is it best at?›

#74 of 200 in Embeddings.

Is Text Embedding 3 Small better than Sgpt Bloom 7b1 Msmarco?›

On MTEB Indic v1 · Bitext Mining · multilingual (20 languages), Text Embedding 3 Small scores 12.2 vs 12.7 — about the same.

How can I use Text Embedding 3 Small?›

Through the OpenAI API.

When was Text Embedding 3 Small released? Is there a newer version?›

Released 2024-01-25. It is the current version in its line.

Is Text Embedding 3 Small open source?›

No, Text Embedding 3 Small is a closed model available through an API. The closest open-weights option here is Multilingual E5 Large Instruct.

What is Text Embedding 3 Small?›

Text Embedding 3 Small is a closed embeddings & retrieval model from OpenAI, released Jan 2024.

What are the best alternatives to Text Embedding 3 Small?›

Multilingual E5 Large Instruct (cheaper).