NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 needs about 9.6 GB of memory at its main file (a 2.0 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 2 quants, from 2.0 GB to 18.5 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 28.5 GB.
NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 – 30B total, 3B active, 1M‑token context
NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 is a hybrid Mamba‑2 + MoE + attention model. It contains 30.6B parameters, with 3B active parameters. The model supports a native context length of 1,048,576 tokens.
Good at
- Large context handling
- Hybrid MoE architecture
- Efficient 1M‑token KV cache
Limits
- No vision support
- Requires 24 GB GPU for full 1M context
- License is OpenMDW‑1.1
Best for
- Long‑form content generation
- Complex reasoning tasks
- Agentic applications
Who it’s for. Developers and researchers building specialized AI agents who need a large‑parameter model with a very long context window. The model is ready for commercial use under the OpenMDW‑1.1 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 2.0 GB | 9.6 GB | Too big | Get file → | |
| — | 18.5 GB | 26.1 GB | Too big | Get file → |
For 8 GB VRAM use the 2 GB file (est. 9.6 GB memory at 8 k context). For 24 GB VRAM use the 18.49 GB file (est. 26.1 GB memory at 8 k context). Memory estimates are approximate.
What GPU do you need for NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4?
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: the best quant that fits is the main file (2.0 GB file, about 9.6 GB in use).
- 16 GB VRAM: the best quant that fits is the main file (2.0 GB file, about 9.6 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (2.0 GB file, about 9.6 GB in use).
- 32 GB VRAM: the best quant that fits is the main file (18.5 GB file, about 26.1 GB in use).
Specifications
Architecture from the GGUF header
Frequently asked questions about NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4
How much VRAM does NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 need?
About 9.6 GB at its main file with an 8K context window. That is the 2.0 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 NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 9.6 GB, so you would need a larger GPU or CPU offloading.
What is the file size of NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 GGUF?
It depends on the quant: file is 2.0 GB, file is 18.5 GB.
What context length does NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 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 7.1 GB at 8K and 28.5 GB at 32K.
Where can I download NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 GGUF?
The files are on Hugging Face in the tngtech/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4-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 NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 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 this model on 8 GB VRAM?
Only the 2 GB quantized file can be loaded, but it will require about 9.6 GB of VRAM at 8 k context, so 8 GB is insufficient.
Does this model support vision?
No, the model has no vision capabilities.
What context length does it support?
The model supports a native context length of 1,048,576 tokens.