Nemotron 3 Embed 8B
Hugging Face (open weights) Open weights
The catch: Large for an embedding model (8B), so it needs a GPU.
| Access | Open weights |
|---|---|
| Free limits | Free to download and self-host; you pay only for your own compute. 8B multilingual text embedding model; 4096-dimension vectors (can be sliced to 2048 or 1024); 32,768-token input. Smaller sibling: nvidia/Nemotron-3-Embed-1B-BF16. Released 14 Jul 2026. |
| Modality | embeddings |
| Credit card | Not required |
| Commercial use | Allowed |
| Trains on your data | – |
| Licence or terms | OpenMDW-1.1 |
| Context window | 32,768 tokens |
| Model id | nvidia/Nemotron-3-Embed-8B-BF16 |
| Base URL | – |
| Last checked | 8 Oct 2026 |
How to use Nemotron 3 Embed 8B
Download the weights and serve them yourself; the quickstart below shows one way. Licence: OpenMDW-1.1.
Quickstart
pip install -U "transformers>=5.2.0" "sentence-transformers>=5.4.1" # then SentenceTransformer("nvidia/Nemotron-3-Embed-8B-BF16"), as in the model cardPaste into Claude Code
I want to run the open-weight model "Nemotron 3 Embed 8B" (nvidia/Nemotron-3-Embed-8B-BF16) for my project. Please: 1. Check the model card and tell me the hardware it needs and the simplest way to serve it (vLLM, SGLang, Ollama or a hosted endpoint). 2. Give me a minimal working setup and a test request. 3. Tell me the licence terms for commercial use. 4. If a cheaper or smaller variant would do the job, mention it.
Before you ship on it
- Large for an embedding model (8B), so it needs a GPU.
- OpenMDW License Agreement v1.1 (linked from the model card): permission to deal in the model materials without restriction, including commercially; if you distribute it you must include the agreement and keep notices;…
- Free to run means you pay for the GPUs, the serving stack and the on-call.
Questions
Is Nemotron 3 Embed 8B free?
Yes. Open weights offers it as open weights, with these limits: Self-host free · OpenMDW-1.1. No credit card is needed.
Can I use Nemotron 3 Embed 8B commercially?
Yes, under OpenMDW-1.1. Check the current terms before you ship.
How do I start with Nemotron 3 Embed 8B?
Download the weights and serve them yourself; the quickstart below shows one way. Licence: OpenMDW-1.1.
Source: huggingface.co/nvidia/Nemotron-3-Embed-8B-BF16, checked 8 Oct 2026.