z

Zembed 1

Mar 2026
zZembed 1
Compare in
on
$0.05per 1M tokens
60.1MTEB Scandinavian v1
Mar 2026released
2,560 (or 1,280, 640)Dimensions
33K tokensMax input

Ranking?

Full ranking →
1Gte Qwen2 7B Instruct80.12Seed1.6 Embedding 121579.33Qwen 3 Embedding 8B79.24Gte Qwen1.5 7B Instruct77.95Octen Embedding 8B77.1
···
37Voyage Large 2 Instruct64.338Zembed 164.239Harrier OSS v1 0.6B64.1

Alternatives?

ModelScoreCostPriceWhat changes
CheaperQwen 3 Embedding 8BunstatedOpen+24%-80%$0.01Memory (4-bit): 2.4 GB → 4.5 GBOpenGte Qwen2 7B Instructunstated+25%——Memory (4-bit): 2.4 GB → 4.2 GBCheaperQwen 3 Embedding 4BunstatedOpen+13%-60%$0.02CheaperQwen 3 Embedding 0.6BunstatedOpen+5%-80%$0.01CheaperJina Embeddings v5 Text NanounstatedOpen+4%-60%$0.02Max input: 33K tokens → 8K tokens

Run it yourself

estimate
Fits on✓ 8 GB (phone, laptop) · 8-bit✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceCC BY-NC 4.0free for commercial use

Benchmarks

in Embeddings & Retrieval
MTEB Europe v1 · Bitext Mining · multilingual (28 languages)80.1#80
MTEB Europe v1 · Clustering · multilingual (29 languages)38.8#69
MTEB Europe v1 · Instruction Reranking · English4.7#15
MTEB Europe v1 · Reranking · multilingual (12 languages)62.1#13
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)17.1#46
MTEB French v1 · Pair Classification · French64.2#38
MTEB Indic v1 · Reranking · multilingual (2 languages)86.2#31
MTEB Indic v1 · Pair Classification · Hindi63.1#83
MTEB Indic v1 · STS · multilingual (13 languages)36.7#57
MTEB Indic v1 · Clustering · multilingual (21 languages)23.2#83
MTEB Dutch v1 · Reranking · Dutch90.7#24
MTEB Portuguese v1 · Multilabel Classification · Portuguese18#143
FollowIR · Mean over tasks · English4.7#15
FollowIR · Mean over task types · English4.7#15
FollowIR · Instruction Reranking · English4.7#15
MTEB Spanish v1 · Pair Classification · Spanish64.2#66
MTEB Thai v1 · Clustering · Thai26.8#99
MTEB Thai v1 · Pair Classification · Thai64.3#84
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)60.1#141
MTEB Russian v1.1 · Pair Classification · Russian53.4#103
MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages)56.6#47
MTEB multilingual v2 · Mean over task types · multilingual (1037 languages)50#43
MTEB multilingual v2 · Classification · multilingual (119 languages)58.1#51
MTEB multilingual v2 · Clustering · multilingual (208 languages)41.8#71
MTEB multilingual v2 · STS · multilingual (35 languages)63.4#69
MTEB multilingual v2 · Pair Classification · multilingual (24 languages)75.8#79
MTEB multilingual v2 · Retrieval · multilingual (120 languages)63.4#27
MTEB multilingual v2 · Bitext Mining · multilingual (1026 languages)57.9#75
MTEB multilingual v2 · Instruction Reranking · English4.7#15
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)18.8#73
MTEB multilingual v2 · Reranking · multilingual (22 languages)65.9#14

FAQ

How much does Zembed 1 cost?›

$0.05 per 1M tokens.

Is Zembed 1 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 Zembed 1, and what is it best at?›

#135 of 205 in Embeddings.

Is Zembed 1 better than Voyage Large 2 Instruct?›

On MTEB French v1 · Pair Classification · French, Zembed 1 scores 64.2 vs 64.3 — about the same. It is 58% cheaper.

When was Zembed 1 released? Is there a newer version?›

Released 2026-03-02. It is the current version in its line.

Is Zembed 1 open source?›

Yes, the weights are open (cc-by-nc-4.0 licence).

Can I run Zembed 1 locally?›

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

What is Zembed 1?›

Zembed 1 is an open-weights embeddings & retrieval model from zeroentropy, released Mar 2026.

What are the best alternatives to Zembed 1?›

Qwen 3 Embedding 8B (cheaper), Gte Qwen2 7B Instruct (open weights).