KURE v3 measured on 4 of 194 Embeddings benchmarks4
BenchmarkByScoreRank
Retrieval1
Other3
No benchmark matches
price not verifiedprice
62.1RTEB French beta
Oct 2026released
8K tokensMax input
en, koLanguages
Alternatives?
ModelScoreCostPriceWhat changes
nOpenKURE v2same——nOpenKURE Sparse Basesame——sOpenComsat Embed Ko 8B Previewsame——Memory (4-bit): 0.7 GB → 4.5 GB
OpenQwen 3 Embedding 8Bunstated-2%—$0.01Languages: not verified+1OpenQwen 3 Embedding 4Bunstated-4%—$0.02Languages: not verified+1ModelScoreCostPriceWhat changes
Nothing here yet.
ModelScoreCostPriceWhat changes
Nothing here yet.
Run it yourself
estimateFits 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
History
3 versionsBenchmarks
in Embeddings & RetrievalFAQ
How much does KURE v3 cost?›
There is no per-unit API price: price not verified.
How good is KURE v3, and what is it best at?›
#5 of 215 in Embeddings.
Is KURE v3 better than KURE v2?›
On MTEB Korean v2 · Retrieval · multilingual (2 languages), KURE v3 scores 80.3 vs 81.1 — about the same.
When was KURE v3 released? Is there a newer version?›
Released 2026-10-20. It is the current version in its line.
Is KURE v3 open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run KURE v3 locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
What is KURE v3?›
KURE v3 is an open-weights embeddings & retrieval model from nlpai-lab, released Oct 2026.
What are the best alternatives to KURE v3?›
KURE v2 (open weights).