Nemotron 3 Embed 8B BF16
Jul 2026
Nemotron 3 Embed 8B BF16Compare inon
Nemotron 3 Embed 8B BF16 measured on 14 of 261 Embeddings benchmarks14
BenchmarkByScoreRank
Reranking9
Other5Leads on 2
No benchmark matches
self-host onlyprice
64.8RTEB French beta
Jul 2026released
4,096Dimensions
33K tokensMax input
en, es, fr +30Languages
Alternatives?
ModelScoreCostPriceWhat changes
BetterEmbeddingGemma 300MunstatedOpen+7%—$0.002Max input: 33K tokens → 2K tokens+1OpenLlama Embed Nemotron 8B+18%——Languages: not verified
BetterQwen 3 Embedding 8BunstatedOpen+17%—$0.01Languages: not verifiediBetterMultilingual E5 Large InstructunstatedOpen+10%—$0.01Max input: 33K tokens → 1K tokensiBetterMultilingual E5 LargeunstatedOpen+4%—$0.01Max input: 33K tokens → 1K tokensModelScoreCostPriceWhat changes
Nothing here yet.
ModelScoreCostPriceWhat changes
EmbeddingGemma 300MunstatedOpen+7%—$0.002Max input: 33K tokens → 2K tokens+1Qwen 3 Embedding 8BunstatedOpen+17%—$0.01Languages: not verifiediMultilingual E5 Large InstructunstatedOpen+10%—$0.01Max input: 33K tokens → 1K tokensiMultilingual E5 LargeunstatedOpen+4%—$0.01Max input: 33K tokens → 1K tokensQwen 3 Embedding 4BunstatedOpen+15%—$0.02Languages: not verifiedModelScoreCostPriceWhat changes
Llama Embed Nemotron 8BOpen+18%——Languages: not verifiedOOcten Embedding 8BOpen+17%——Languages: loses es, fr, de +28Qwen 3 Embedding 8BunstatedOpen+17%—$0.01Languages: not verifiedcF2LLM v2 14BunstatedOpen+17%——Memory (4-bit): 4.8 GB → 8.4 GBcF2LLM v2 8BunstatedOpen+16%——Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · compressed✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
Licenceopenmdw-1.1commercial use with conditions
Benchmarks
in Embeddings & RetrievalMTEB Dutch v1 · STS · Dutch71.4#62
FAQ
How much does Nemotron 3 Embed 8B BF16 cost?›
There is no per-unit API price: self-host only.
Is Nemotron 3 Embed 8B BF16 free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is Nemotron 3 Embed 8B BF16, and what is it best at?›
#2 of 212 in Embeddings.
Is Nemotron 3 Embed 8B BF16 better than Granite Embedding 311M Multilingual R2?›
On MTEB Dutch v1 · Mean over task types · Dutch, Nemotron 3 Embed 8B BF16 scores 59.8 vs 59.9 — about the same.
When was Nemotron 3 Embed 8B BF16 released? Is there a newer version?›
Released 2026-07-16. It is the current version in its line.
Is Nemotron 3 Embed 8B BF16 open source?›
Yes, the weights are open (openmdw-1.1 licence).
Can I run Nemotron 3 Embed 8B BF16 locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
What is Nemotron 3 Embed 8B BF16?›
Nemotron 3 Embed 8B BF16 is an open-weights embeddings & retrieval model from NVIDIA, released Jul 2026.
What are the best alternatives to Nemotron 3 Embed 8B BF16?›
EmbeddingGemma 300M (better), Llama Embed Nemotron 8B (open weights).