KaLM Embedding Multilingual Mini Instruct v2 measured on 15 of 261 Embeddings benchmarks15
BenchmarkByScoreRank
Retrieval8
Reranking1
Other6
No benchmark matches
self-host onlyprice
86.2MTEB English v2
Jun 2025released
896Dimensions
33K tokensMax input
en, zhLanguages
Alternatives?
ModelScoreCostPriceWhat changes
iBetterE5 Mistral 7B InstructunstatedOpen+3%—$0.022Languages: loses zh+1KOpenQZhou Embeddingunstated+7%——Memory (4-bit): 0.3 GB → 4.2 GBVBetterVoyage Large 2 Instructunstated+6%—$0.12Max input: 33K tokens → 16K tokens+1jBetterIngot 8B R3unstated+9%——Languages: not verifiedTBetterConan Embedding v2unstated+8%——ModelScoreCostPriceWhat changes
Nothing here yet.
ModelScoreCostPriceWhat changes
iE5 Mistral 7B InstructunstatedOpen+3%—$0.022Languages: loses zh+1VVoyage Large 2 Instructunstated+6%—$0.12Max input: 33K tokens → 16K tokens+1jIngot 8B R3unstated+9%——Languages: not verifiedTConan Embedding v2unstated+8%——KQZhou EmbeddingunstatedOpen+7%——Memory (4-bit): 0.3 GB → 4.2 GBModelScoreCostPriceWhat changes
KQZhou EmbeddingunstatedOpen+7%——Memory (4-bit): 0.3 GB → 4.2 GBiJasper Token Compression 600MunstatedOpen+5%——Memory (4-bit): 0.3 GB → 0.4 GBKNano Em1 0.6B v2.1unstatedOpen+4%——Max input: 33K tokens → 8K tokens+2
NV Embed v2unstatedOpen+3%——Languages: loses zh+1aLGAI Embedding PreviewunstatedOpen+3%——Languages: loses zh+1ModelScoreCostPriceWhat changes
HKaLM Embedding Multilingual Mini v1unstatedOpen-2%——Max input: 33K tokens → 1K tokens+1HKaLM Embedding Multilingual Mini Instruct v1unstatedOpen-3%——Languages: not verifiedHKaLM Embedding Multilingual Mini Instruct v1.5unstatedOpen-3%——Languages: not verifiedAll HIT-TMG models →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
LicenceMITfree for commercial use
History
3 versionsBenchmarks
in Embeddings & RetrievalBEIR · Retrieval · English51.7#82
FAQ
How much does KaLM Embedding Multilingual Mini Instruct v2 cost?›
There is no per-unit API price: self-host only.
Is KaLM Embedding Multilingual Mini Instruct v2 free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is KaLM Embedding Multilingual Mini Instruct v2, and what is it best at?›
#48 of 238 in Embeddings.
Is KaLM Embedding Multilingual Mini Instruct v2 better than GIST Embedding v0?›
On MTEB English v2 · Pair Classification · English, KaLM Embedding Multilingual Mini Instruct v2 scores 86.2 vs 86.3 — about the same.
When was KaLM Embedding Multilingual Mini Instruct v2 released? Is there a newer version?›
Released 2025-06-25. It is the current version in its line.
Is KaLM Embedding Multilingual Mini Instruct v2 open source?›
Yes, the weights are open (MIT licence).
Can I run KaLM Embedding Multilingual Mini Instruct v2 locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
What is KaLM Embedding Multilingual Mini Instruct v2?›
KaLM Embedding Multilingual Mini Instruct v2 is an open-weights embeddings & retrieval model from HIT-TMG, released Jun 2025.
What are the best alternatives to KaLM Embedding Multilingual Mini Instruct v2?›
E5 Mistral 7B Instruct (better), QZhou Embedding (open weights).