i

Jasper Token Compression 600M

Nov 2025
iJasper Token Compression 600M
Compare in
on
self-host onlyprice
66.2MTEB English v2
Nov 2025released
1,024Dimensions
33K tokensMax input
en, zhLanguages

Rankingconfiguration unstated?

Full ranking →
1Nemotron 3 Embed 8B BF1677.22Dinghy Law 8B v177.13Voyage 4 Large76.94Most Embed De76.25Octen Embedding 8B74.1
···
93Granite Embedding 97M Multilingual R234.394Jasper Token Compression 600M34.395Speed Embedding 7B Instruct33.2

Strong at

S
Embeddings & Retrieval#8 of 200

Alternatives?

ModelScoreCostPriceWhat changes
BetterQwen 3 Embedding 8BunstatedOpen+95%—$0.01Languages: not verified+1OpenNemotron 3 Embed 8B BF16×2.3——Memory (4-bit): 0.4 GB → 4.8 GBiBetterMultilingual E5 Large InstructunstatedOpen+49%—$0.01Max input: 33K tokens → 1K tokensBetterBge M3unstatedOpen+47%—$0.01Max input: 33K tokens → 8K tokens+1iBetterMultilingual E5 LargeunstatedOpen+26%—$0.01Max input: 33K tokens → 1K 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
MTEB Mandarin Chinese v1 · Mean over tasks · Mandarin Chinese73.5#8
MTEB Mandarin Chinese v1 · Mean over task types · Mandarin Chinese75#7
MTEB Mandarin Chinese v1 · Retrieval · Mandarin Chinese75.6#20
MTEB Mandarin Chinese v1 · STS · Mandarin Chinese63.9#10
MTEB Mandarin Chinese v1 · Reranking · Mandarin Chinese70#14
MTEB Mandarin Chinese v1 · Clustering · Mandarin Chinese77.5#5
MTEB Mandarin Chinese v1 · Pair Classification · Mandarin Chinese85.4#12
MTEB Mandarin Chinese v1 · Classification · Mandarin Chinese77.9#6
MTEB Law v1 · Mean over tasks · multilingual (3 languages)55.4#34
MTEB Law v1 · Mean over task types · multilingual (3 languages)55.4#34
MTEB Law v1 · Retrieval · multilingual (3 languages)55.4#34
MTEB Medical v1 · Clustering · English47.6#20
MTEB Medical v1 · Reranking · Mandarin Chinese88.6#18
MTEB Thai v1 · Mean over tasks · multilingual (2 languages)56.5#23
MTEB Thai v1 · Mean over task types · multilingual (2 languages)53.8#23
MTEB Thai v1 · Clustering · Thai37.8#86
MTEB Thai v1 · Classification · Thai62.4#23
MTEB Thai v1 · Pair Classification · Thai64.8#82
MTEB Thai v1 · Retrieval · multilingual (2 languages)57.6#22
MTEB Thai v1 · Reranking · Thai46.5#23
RTEB German beta · Mean over tasks · German34.3#94
RTEB Law beta · Mean over tasks · multilingual (4 languages)43.9#70

FAQ

How much does Jasper Token Compression 600M cost?›

There is no per-unit API price: self-host only.

Is Jasper Token Compression 600M free to use?›

There is no free API. The weights are open, so you can run it yourself.

How good is Jasper Token Compression 600M, and what is it best at?›

#8 of 200 in Embeddings. Strongest at: Embeddings & Retrieval.

Is Jasper Token Compression 600M better than Granite Embedding 97M Multilingual R2?›

On RTEB German beta · Mean over tasks · German, Jasper Token Compression 600M scores 34.3 vs 34.3 — about the same.

When was Jasper Token Compression 600M released? Is there a newer version?›

Released 2025-11-14. It is the current version in its line.

Is Jasper Token Compression 600M open source?›

Yes, the weights are open (MIT licence).

Can I run Jasper Token Compression 600M locally?›

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

What is Jasper Token Compression 600M?›

Jasper Token Compression 600M is an open-weights embeddings & retrieval model from infgrad, released Nov 2025.

What are the best alternatives to Jasper Token Compression 600M?›

Qwen 3 Embedding 8B (better), Nemotron 3 Embed 8B BF16 (open weights).