Jina Embeddings v2 Base En

Sep 2023
Jina Embeddings v2 Base En
Compare in
on
$0.05per 1M tokens
26.7MTEB English v2
Sep 2023released
768Dimensions
8K tokensMax input
enLanguages

Rankingconfiguration unstated?

Full ranking →
1Uibk Embedding v191.52GTE ModernColBERT v190.63UltraBERT Embed 150M79.34Voyage Multilingual 279.25Voyage Law 278.8
···
170Deberta v1 Base4.4171Jina Embeddings v2 Base En1.8

Strong at

C
Embeddings & Retrieval#169 of 200

Alternatives?

ModelScoreCostPriceWhat changes
VCheaperVoyage 3 Liteunstated×39.3-60%$0.02Languages: not verifiediBetterE5 Base v2unstatedOpen×22.6-90%$0.005Max input: 8K tokens → 1K tokensDOpenUibk Embedding v1unstated×50.4——Max input: 8K tokens → 0K tokensCheaperJina Embeddings v5 Text NanounstatedOpen×35-60%$0.02Languages: not verifiedCheaperBge M3unstatedOpen×32.7-80%$0.01Languages: not verified+1

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 · English45.3#129
BEIR · Mean over task types · English45.3#129
BEIR · Retrieval · English45.3#129
JMTEB v2 · Clustering · Japanese12.7#126
JMTEB v2 · STS · Japanese27.7#151
LongEmbed · Mean over tasks · English1.8#171
LongEmbed · Mean over task types · English1.8#171
LongEmbed · Retrieval · English1.8#171
MTEB German v1 · Classification · German19.3#165
MTEB German v1 · Pair Classification · German46.7#169
MTEB Europe v1 · Retrieval · multilingual (25 languages)8.6#137
MTEB Europe v1 · Bitext Mining · multilingual (28 languages)1.3#179
MTEB Europe v1 · Classification · multilingual (23 languages)35.3#135
MTEB Europe v1 · Clustering · multilingual (29 languages)12#170
MTEB Europe v1 · Instruction Reranking · English-1.6#110
MTEB Europe v1 · Pair Classification · multilingual (11 languages)68.5#140
MTEB Europe v1 · Reranking · multilingual (12 languages)35.7#163
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)0.8#176
MTEB French v1 · Reranking · French29.4#137
MTEB French v1 · Classification · French13.7#131
MTEB French v1 · Summarization · French27.7#144
MTEB French v1 · Pair Classification · French50#179
MTEB Indic v1 · Mean over tasks · multilingual (25 languages)28.9#121
MTEB Indic v1 · Mean over task types · multilingual (25 languages)20.9#113
MTEB Indic v1 · Classification · multilingual (8 languages)39#134
MTEB Indic v1 · Retrieval · multilingual (14 languages)1.7#148
MTEB Indic v1 · Bitext Mining · multilingual (20 languages)0.2#158
MTEB Indic v1 · Reranking · multilingual (2 languages)54.6#116
MTEB Indic v1 · Pair Classification · Hindi50.9#168
MTEB Indic v1 · STS · multilingual (13 languages)-2.1#183
MTEB Indic v1 · Clustering · multilingual (21 languages)1.7#189
MTEB Dutch v1 · Reranking · Dutch74.9#158
MTEB Polish v1 · Pair Classification · Polish41#153
MTEB Polish v1 · Classification · Polish27.3#137
MTEB Portuguese v1 · Multilabel Classification · Portuguese17.6#155
MTEB Law v1 · Mean over tasks · multilingual (3 languages)3.2#163
MTEB Law v1 · Mean over task types · multilingual (3 languages)3.2#163
MTEB Law v1 · Retrieval · multilingual (3 languages)3.2#163
MTEB Medical v1 · Mean over tasks · multilingual (10 languages)21.1#121
MTEB Medical v1 · Mean over task types · multilingual (10 languages)15.6#125
MTEB Medical v1 · Retrieval · multilingual (10 languages)23.8#115
MTEB Medical v1 · Clustering · English16#231
MTEB Medical v1 · Reranking · Mandarin Chinese7#198
FollowIR · Mean over tasks · English-1.6#110
FollowIR · Mean over task types · English-1.6#110
FollowIR · Instruction Reranking · English-1.6#110
MTEB Spanish v1 · Pair Classification · Spanish49.4#174
MTEB Thai v1 · Clustering · Thai2.3#188
MTEB Thai v1 · Pair Classification · Thai50.5#184
MTEB Scandinavian v1 · Clustering · multilingual (2 languages)14.9#171
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)4.3#211
MTEB Russian v1.1 · Pair Classification · Russian44.8#191
MTEB Russian v1.1 · Multilabel Classification · Russian23.6#159
RTEB Finnish beta · Mean over tasks · English52.4#65
RTEB French beta · Mean over tasks · multilingual (2 languages)0.1#217
RTEB German beta · Mean over tasks · German2.6#211
RTEB Health beta · Mean over tasks · multilingual (4 languages)27.3#79
RTEB Law beta · Mean over tasks · multilingual (4 languages)4.5#210
MTEB multilingual v2 · Classification · multilingual (119 languages)34.2#157
MTEB multilingual v2 · Clustering · multilingual (208 languages)21.1#163
MTEB multilingual v2 · Pair Classification · multilingual (24 languages)68.2#167
MTEB multilingual v2 · Retrieval · multilingual (120 languages)10.7#172
MTEB multilingual v2 · Bitext Mining · multilingual (1026 languages)0.9#177
MTEB multilingual v2 · Instruction Reranking · English-1.6#110
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)10.4#176
MTEB multilingual v2 · Reranking · multilingual (22 languages)31.6#172

FAQ

How much does Jina Embeddings v2 Base En cost?›

$0.05 per 1M tokens.

Is Jina Embeddings v2 Base En free to use?›

There is no free API; it costs $0.05 per 1M tokens. The weights are open, so you can run it yourself.

How good is Jina Embeddings v2 Base En, and what is it best at?›

#169 of 200 in Embeddings.

Is Jina Embeddings v2 Base En better than Deberta v1 Base?›

On LongEmbed · Mean over task types · English, Jina Embeddings v2 Base En scores 1.8 vs 4.4 — lower.

When was Jina Embeddings v2 Base En released? Is there a newer version?›

Released 2023-09-27. It is the current version in its line.

Is Jina Embeddings v2 Base En open source?›

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

Can I run Jina Embeddings v2 Base En locally?›

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

What is Jina Embeddings v2 Base En?›

Jina Embeddings v2 Base En is an open-weights embeddings & retrieval model from Jina AI, released Sep 2023.

What are the best alternatives to Jina Embeddings v2 Base En?›

Voyage 3 Lite (cheaper), E5 Base v2 (better), Uibk Embedding v1 (open weights).