h

Bekko Static Embedding v1

Oct 2026
hBekko Static Embedding v1
Compare in
on
price not verifiedprice
60.7MTEB Scandinavian v1
Oct 2026released
33K tokensMax input

Ranking?

Full ranking →
1Harrier OSS v1 27B79.92KaLM Embedding Gemma3 12B 251177.93Seed1.6 Embedding 121576.74Qwen 3 Embedding 8B745Harrier OSS v1 0.6B73.9
···
73Granite Embedding 107M Multilingual52.674Bekko Static Embedding v152.475Potion Multilingual 128M52.4

Alternatives?

ModelScoreCostPriceWhat changes
BetterEmbeddingGemma 300MunstatedOpen+16%—$0.002Max input: 33K tokens → 2K tokens+1OpenKaLM Embedding Gemma3 12B 2511+49%——Memory (4-bit): 0.1 GB → 7.1 GBBetterQwen 3 Embedding 8BunstatedOpen+41%—$0.01Memory (4-bit): 0.1 GB → 4.5 GBBetterQwen 3 Embedding 0.6BunstatedOpen+28%—$0.01Memory (4-bit): 0.1 GB → 0.4 GBiBetterMultilingual E5 Large InstructunstatedOpen+24%—$0.01Max input: 33K tokens → 1K tokens+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
LicenceMITfree for commercial use

Benchmarks

in Embeddings & Retrieval
MTEB Europe v1 · Instruction Reranking · English-1.7#111
MTEB Europe v1 · Multilabel Classification · multilingual (23 languages)3.9#169
MTEB Europe v1 · Clustering · multilingual (29 languages)37.1#85
MTEB Europe v1 · Reranking · multilingual (12 languages)48.8#119
MTEB Europe v1 · Bitext Mining · multilingual (28 languages)65.6#93
MTEB French v1 · Pair Classification · French47.3#192
MTEB Indic v1 · STS · multilingual (13 languages)12.5#94
MTEB Indic v1 · Reranking · multilingual (2 languages)80.5#76
MTEB Indic v1 · Pair Classification · Hindi61.8#89
MTEB Indic v1 · Clustering · multilingual (21 languages)7.5#105
MTEB Dutch v1 · Reranking · Dutch79.9#119
MTEB Portuguese v1 · Multilabel Classification · Portuguese20.6#65
FollowIR · Mean over tasks · English-1.7#111
FollowIR · Mean over task types · English-1.7#111
FollowIR · Instruction Reranking · English-1.7#111
MTEB Spanish v1 · Pair Classification · Spanish54.2#163
MTEB Thai v1 · Pair Classification · Thai60.8#98
MTEB Thai v1 · Clustering · Thai27.2#98
MTEB Scandinavian v1 · Bitext Mining · multilingual (3 languages)60.7#136
MTEB Russian v1.1 · Pair Classification · Russian52.8#110
MTEB multilingual v2 · Mean over tasks · multilingual (1037 languages)50.2#56
MTEB multilingual v2 · Mean over task types · multilingual (1037 languages)43.1#56
MTEB multilingual v2 · Clustering · multilingual (208 languages)40#100
MTEB multilingual v2 · STS · multilingual (35 languages)64.5#68
MTEB multilingual v2 · Classification · multilingual (119 languages)52.4#74
MTEB multilingual v2 · Retrieval · multilingual (120 languages)50#81
MTEB multilingual v2 · Multilabel Classification · multilingual (25 languages)15#138
MTEB multilingual v2 · Bitext Mining · multilingual (1026 languages)43.7#89
MTEB multilingual v2 · Instruction Reranking · English-1.7#111
MTEB multilingual v2 · Reranking · multilingual (22 languages)51.6#80
MTEB multilingual v2 · Pair Classification · multilingual (24 languages)72.8#104

FAQ

How much does Bekko Static Embedding v1 cost?›

There is no per-unit API price: price not verified.

How good is Bekko Static Embedding v1, and what is it best at?›

#130 of 206 in Embeddings.

Is Bekko Static Embedding v1 better than Granite Embedding 107M Multilingual?›

On MTEB multilingual v2 · Classification · multilingual (119 languages), Bekko Static Embedding v1 scores 52.4 vs 52.6 — about the same.

When was Bekko Static Embedding v1 released? Is there a newer version?›

Released 2026-10-21. It is the current version in its line.

Is Bekko Static Embedding v1 open source?›

Yes, the weights are open (MIT licence).

Can I run Bekko Static Embedding v1 locally?›

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

What is Bekko Static Embedding v1?›

Bekko Static Embedding v1 is an open-weights embeddings & retrieval model from hotchpotch, released Oct 2026.

What are the best alternatives to Bekko Static Embedding v1?›

EmbeddingGemma 300M (better), KaLM Embedding Gemma3 12B 2511 (open weights).