Bge M3

Jun 2024
Bge M3
Compare in
on
$0.01per 1M tokens
32.2MTEB Medical v1
Jun 2024released
1,024Dimensions
8K tokensMax input

Rankingconfiguration unstated?

Full ranking →
1Gemini Embedding 001832Jina Embeddings v5 Omni Small82.53Jina Embeddings v5 Text Small82.54Jina Embeddings v5 Omni Nano81.74Jina Embeddings v5 Text Nano81.7
···
21Multilingual E5 Base78.322Bge M37823SFR Embedding 2 R77.8

Alternatives?

ModelScoreCostPriceWhat changes
iCheaperE5 Base v2unstatedOpen-12%-50%$0.005Max input: 8K tokens → 1K tokensiBetterMultilingual E5 Large InstructunstatedOpen+3%same$0.01Max input: 8K tokens → 1K tokensOpenJina Embeddings v5 Text Smallunstated+6%+400%$0.05Memory (4-bit): 0.3 GB → 0.4 GBBetterJina Embeddings v5 Text NanounstatedOpen+5%+100%$0.02BetterGemini Embedding 001unstated+6%+1,400%$0.15Max input: 8K tokens → 2K tokens

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
LicenceMITfree for commercial use

Benchmarks

in Embeddings & Retrieval
BEIR-NL · Mean over tasks · Dutch44.7#5
BEIR-NL · Mean over task types · Dutch44.7#5
BEIR-NL · Retrieval · Dutch44.7#5
ChemTEB · Mean over tasks · multilingual (13 languages)69.6#27
ChemTEB · Mean over task types · multilingual (13 languages)50.3#23
ChemTEB · Classification · English80.7#30
ChemTEB · Retrieval · English70.5#27
ChemTEB · Pair Classification · multilingual (13 languages)55.4#32
ChemTEB · Clustering · English44.7#24
ChemTEB · Bitext Mining · English0.1#44
MTEB Persian v2 · Mean over tasks · Persian65.6#10
MTEB Persian v2 · Mean over task types · Persian68.6#10
MTEB Persian v2 · STS · Persian79.7#6
MTEB Persian v2 · Clustering · Persian57.7#29
MTEB Persian v2 · Classification · Persian65.3#19
MTEB Persian v2 · Retrieval · Persian58.9#6
MTEB Persian v2 · Reranking · Persian74.6#2
MTEB Persian v2 · Bitext Mining · Persian61.1#16
MTEB Persian v2 · Pair Classification · Persian83.2#15
JMTEB v2 · Clustering · Japanese42.6#60
JMTEB v2 · STS · Japanese79.7#45
LongEmbed · Mean over tasks · English59.5#37
LongEmbed · Mean over task types · English59.5#37
LongEmbed · Retrieval · English59.5#37
MTEB German v1 · Classification · German68.9#26
MTEB Europe v1 · Clustering · multilingual (29 languages)36.9#87
MTEB Europe v1 · Instruction Reranking · English-3.1#169
MTEB Europe v1 · Reranking · multilingual (12 languages)57.5#57
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)13.1#90
MTEB French v1 · Pair Classification · French59.2#78
MTEB Indic v1 · Reranking · multilingual (2 languages)86#32
MTEB Indic v1 · Pair Classification · Hindi77.5#28
MTEB Indic v1 · STS · multilingual (13 languages)52.1#24
MTEB Indic v1 · Clustering · multilingual (21 languages)30#72
MTEB Dutch v1 · Mean over tasks · Dutch54.8#34
MTEB Dutch v1 · Mean over task types · Dutch61.6#31
MTEB Dutch v1 · Classification · Dutch61.2#34
MTEB Dutch v1 · Clustering · Dutch28.7#58
MTEB Dutch v1 · Retrieval · Dutch60#25
MTEB Dutch v1 · STS · Dutch79#33
MTEB Dutch v1 · Multilabel Classification · Dutch35.3#43
MTEB Dutch v1 · Pair Classification · Dutch78.3#30
MTEB Dutch v1 · Reranking · Dutch88.7#49
MTEB Portuguese v1 · Multilabel Classification · Portuguese19.8#90
MTEB Medical v1 · Clustering · English32.2#174
MTEB Medical v1 · Reranking · Mandarin Chinese79.2#51
FollowIR · Mean over tasks · English-3.1#169
FollowIR · Mean over task types · English-3.1#169
FollowIR · Instruction Reranking · English-3.1#169
MTEB Spanish v1 · Mean over tasks · multilingual (2 languages)64.8#14
MTEB Spanish v1 · Mean over task types · multilingual (2 languages)66.3#12
MTEB Spanish v1 · Pair Classification · Spanish69.3#39
MTEB Spanish v1 · STS · multilingual (2 languages)78#22
MTEB Spanish v1 · Retrieval · multilingual (2 languages)49.4#16
MTEB Spanish v1 · Clustering · Spanish45.4#36
MTEB Spanish v1 · Reranking · Spanish64#5
MTEB Spanish v1 · Classification · Spanish91.8#8
MTEB Thai v1 · Mean over tasks · multilingual (2 languages)64.2#10
MTEB Thai v1 · Mean over task types · multilingual (2 languages)61.8#15
MTEB Thai v1 · Clustering · Thai36.3#93
MTEB Thai v1 · Classification · Thai64.8#16
MTEB Thai v1 · Pair Classification · Thai78.5#26
MTEB Thai v1 · Retrieval · multilingual (2 languages)66.1#5
MTEB Thai v1 · Reranking · Thai63.2#10
MTEB Scandinavian v1 · Mean over tasks · multilingual (6 languages)57.9#22
MTEB Scandinavian v1 · Mean over task types · multilingual (6 languages)58.4#23
MTEB Scandinavian v1 · Clustering · multilingual (2 languages)38.8#55
MTEB Scandinavian v1 · Classification · multilingual (6 languages)61.7#29
MTEB Scandinavian v1 · Retrieval · multilingual (3 languages)64#26
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)69.3#52
VN-MTEB Vietnamese v1 · Mean over tasks · Vietnamese53.6#7
VN-MTEB Vietnamese v1 · Mean over task types · Vietnamese63.5#8
VN-MTEB Vietnamese v1 · Classification · Vietnamese67.4#10
VN-MTEB Vietnamese v1 · STS · Vietnamese78.8#14
VN-MTEB Vietnamese v1 · Retrieval · Vietnamese39.8#8
VN-MTEB Vietnamese v1 · Clustering · Vietnamese45.9#13
VN-MTEB Vietnamese v1 · Reranking · Vietnamese64.7#12
VN-MTEB Vietnamese v1 · Pair Classification · Vietnamese84.6#5
MTEB Russian v1.1 · Mean over tasks · Russian61.6#24
MTEB Russian v1.1 · Mean over task types · Russian61#20
MTEB Russian v1.1 · Classification · Russian60.4#44
MTEB Russian v1.1 · STS · Russian73.7#26
MTEB Russian v1.1 · Reranking · Russian69.7#15
MTEB Russian v1.1 · Clustering · Russian52.4#55
MTEB Russian v1.1 · Pair Classification · Russian60.6#36
MTEB Russian v1.1 · Retrieval · Russian75.2#10
MTEB Russian v1.1 · Multilabel Classification · Russian34.9#49
RTEB Finnish beta · Mean over tasks · English57.7#46
RTEB French beta · Mean over tasks · multilingual (2 languages)50.1#41
RTEB German beta · Mean over tasks · German50.4#60
RTEB Health beta · Mean over tasks · multilingual (4 languages)55#43
RTEB Law beta · Mean over tasks · multilingual (4 languages)44.7#68
MTEB multilingual v2 · Classification · multilingual (119 languages)60.3#39
MTEB multilingual v2 · Clustering · multilingual (208 languages)40.9#83
MTEB multilingual v2 · STS · multilingual (35 languages)74.1#32
MTEB multilingual v2 · Pair Classification · multilingual (24 languages)80.8#31
MTEB multilingual v2 · Retrieval · multilingual (120 languages)54.6#69
MTEB multilingual v2 · Instruction Reranking · English-3.1#169
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)20.1#59
MTEB multilingual v2 · Reranking · multilingual (22 languages)62.8#43

FAQ

How much does Bge M3 cost?›

$0.01 per 1M tokens.

Is Bge M3 free to use?›

There is no free API; the cheapest paid price is $0.01 per 1M input tokens. The weights are open, so you can run it yourself.

How good is Bge M3, and what is it best at?›

#163 of 220 in Embeddings.

Is Bge M3 better than Multilingual E5 Base?›

On MTEB Spanish v1 · STS · multilingual (2 languages), Bge M3 scores 78 vs 78.3 — about the same.

How can I use Bge M3?›

Through the BAAI API.

When was Bge M3 released? Is there a newer version?›

Released 2024-06-28. It is the current version in its line.

Is Bge M3 open source?›

Yes, the weights are open (MIT licence).

Can I run Bge M3 locally?›

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

What is Bge M3?›

Bge M3 is an open-weights embeddings & retrieval model from BAAI, released Jun 2024.

What are the best alternatives to Bge M3?›

E5 Base v2 (cheaper), Multilingual E5 Large Instruct (better), Jina Embeddings v5 Text Small (open weights).