Nemotron-3-Nano-30B-A3B needs about 30.5 GB of memory at its main file (a 23.0 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 7 quants, from 18.8 GB to 27.2 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 27.9 GB.
Nemotron-3-Nano-30B-A3B APEX GGUF – 31.6B parameter MoE model
Nemotron-3-Nano-30B-A3B APEX GGUF is a 31.6B parameter mixture‑of‑experts model from NVIDIA. It uses a hybrid architecture of Mamba‑2, MoE, and GQA attention layers. The model is quantized with APEX and available in several GGUF files.
Good at
- Reasoning with trace generation
- Multilingual support in English, German, Spanish, French, Italian, Japanese
- Efficient MoE inference
Limits
- Requires large VRAM for full 30B model
- No vision capabilities
- License governed by NVIDIA Nemotron Open Model License
Best for
- Chatbot agents needing reasoning traces
- Custom LLM deployment with limited GPU memory
- Research on MoE quantization
Who it’s for. Engineers and researchers building specialized AI agents who need a large, efficient MoE model. The model is ready for commercial use under NVIDIA’s open license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 18.8 GB | 26.2 GB | Too big | Get file → | |
| — | 20.3 GB | 27.8 GB | Too big | Get file → | |
| — | 20.3 GB | 27.8 GB | Too big | Get file → | |
| — | 23.0 GB | 30.5 GB | Too big | Get file → | |
| — | 23.0 GB | 30.5 GB | Too big | Get file → | |
| — | 27.2 GB | 34.7 GB | Too big | Get file → | |
| — | 27.2 GB | 34.7 GB | Too big | Get file → |
For 8‑GB VRAM, the smallest GGUF file (18.77 GB) will still exceed memory; consider using a higher‑compression quant if available. With 16‑GB VRAM, the 20.34 GB file may fit with 8‑k context, estimating 27.8 GB memory usage. For 24‑GB VRAM, the 22.97 GB file is suitable, estimating 30.5 GB memory usage. Memory estimates are approximate.
What GPU do you need for Nemotron-3-Nano-30B-A3B?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: the best quant that fits is the main file (23.0 GB file, about 30.5 GB in use).
- 48 GB VRAM: the best quant that fits is the main file (27.2 GB file, about 34.7 GB in use).
Which Nemotron-3-Nano-30B-A3B quant should you choose?
If memory is tight, the smallest file is the first one in the table at 18.8 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Nemotron-3-Nano-30B-A3B
How much VRAM does Nemotron-3-Nano-30B-A3B need?
About 30.5 GB at its main file with an 8K context window. That is the 23.0 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run Nemotron-3-Nano-30B-A3B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 26.2 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Nemotron-3-Nano-30B-A3B GGUF?
It depends on the quant: file is 18.8 GB, file is 20.3 GB, file is 20.3 GB, file is 23.0 GB, file is 23.0 GB, file is 27.2 GB and 1 more.
What context length does Nemotron-3-Nano-30B-A3B support?
The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 7.0 GB at 8K and 27.9 GB at 32K.
Where can I download Nemotron-3-Nano-30B-A3B GGUF?
The files are on Hugging Face in the mudler/Nemotron-3-Nano-30B-A3B-APEX-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Nemotron-3-Nano-30B-A3B use?
Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Nemotron-3-Nano-30B-A3B on 8 GB VRAM?
No, the smallest GGUF file requires at least 16 GB VRAM; 8 GB is insufficient.
What context length does the model support?
The model was trained on 1,048,576 tokens and supports up to that context length.
Does the model support vision tasks?
No, the model has no vision capabilities.