Skip to content

Update nemotron nano v3 blog for NVFP4 weight #299

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
wisclmy0611 merged 3 commits into from
Jan 28, 2026
Merged

Update nemotron nano v3 blog for NVFP4 weight #299

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
3ddcd65
upd
Fridge003 Jan 20, 2026
c18fc9d
upd install
Fridge003 Jan 20, 2026
6a0b4ff
upd
Fridge003 Jan 28, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 9 additions & 4 deletions blog/2025-12-15-run-nvidia-nemotron-3-nano.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,8 @@ date: "December 15, 2025"
previewImg: /images/blog/nemotron-3-nano/benchmark.png
---

**Jan 28th Update**: NVIDIA just released their Nemotron 3 Nano model in NVFP4 precision. This model is supported by SGLang out of the box and it uses a new method called Quantization-Aware Distillation (QAD) to maintain accuracy on NVFP4 while delivering 4x throughput on B200 compared to FP8-H100. You can download the NVFP4 checkpoints [here](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4) and run them using this [NVIDIA Brev launchable](https://brev.nvidia.com/launchable/deploy?launchableID=env-386BHXsTBKROX8F2WBCbQP6S6qt).

We are excited to announce that SGLang supports the latest highly efficient NVIDIA Nemotron 3 Nano model on Day 0!

Nemotron 3 Nano, part of the newly announced open [Nemotron 3 family](https://developer.nvidia.com/blog/inside-nvidia-nemotron-3-techniques-tools-and-data-that-make-it-efficient-and-accurate/), is a compact MoE language model offering industry-leading compute efficiency and accuracy, enabling developers to build specialized agentic AI systems.
Expand All @@ -28,7 +30,7 @@ Nemotron 3 Nano is fully open with open-weights, datasets and recipes so develop
- Model output: Text
- Supported GPUs: NVIDIA RTX Pro 6000, DGX Spark, H100, B200.
- Get started:
- Download model weights from Hugging Face - [BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16), [FP8](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8)
- Download model weights from Hugging Face - [BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16), [FP8](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8), [NVFP4](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4)
- [Run with SGLang for inference](https://cookbook.sglang.io/docs/NVIDIA/Nemotron3-Nano)
- [Technical report](https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Nano-Technical-Report.pdf) to build custom, optimized models with Nemotron techniques.

Expand All @@ -38,16 +40,19 @@ For an easier setup with SGLang, refer to our getting started cookbook, availabl

Run the command below to install dependencies:
```bash
pip install sglang==0.5.6.post2.dev7852+g8102e36b5 --extra-index-url https://sgl-project.github.io/whl/nightly/
uv pip install sglang==0.5.6.post3.dev1278+gad1b4e472 --extra-index-url https://sgl-project.github.io/whl/nightly/
```

We can then serve this model:
```bash
# BF16
python3 -m sglang.launch_server --model-path nvidia/NVIDIA-Nemotron-Nano-3-30B-A3B-BF16 --trust-remote-code --reasoning-parser nano_v3 --tool-call-parser qwen3_coder

# Swap out model name for FP8
# FP8
python3 -m sglang.launch_server --model-path nvidia/NVIDIA-Nemotron-Nano-3-30B-A3B-FP8 --trust-remote-code --reasoning-parser nano_v3 --tool-call-parser qwen3_coder

# NVFP4
python3 -m sglang.launch_server --model-path nvidia/NVIDIA-Nemotron-Nano-3-30B-A3B-NVFP4 --trust-remote-code --reasoning-parser nano_v3 --tool-call-parser qwen3_coder
```

Once the server is up and running, you can prompt the model using the below code snippets:
Expand Down Expand Up @@ -93,7 +98,7 @@ Trained on NVIDIA-curated, high-quality data, Nemotron 3 Nano leads on benchmark

## Get Started

- Download Nemotron 3 Nano model weights from Hugging Face - [BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16), [FP8](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8)
- Download Nemotron 3 Nano model weights from Hugging Face - [BF16](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16), [FP8](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8), [NVFP4](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4)
- Run with SGLang for inference using [this](https://cookbook.sglang.io/docs/NVIDIA/Nemotron3-Nano) cookbook or through this NVIDIA Brev [launchable](https://brev.nvidia.com/launchable/deploy?launchableID=env-36ikQZX0ZDTSCGE7YkqxiOKwKsj).


Expand Down