Nvidia-Nemotron-3.5-Lightning-30B-A3B needs about 32.9 GB of memory at Q4_K_M (a 25.3 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 18 quants, from 19.4 GB (IQ1_M) to 65.9 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 28.5 GB.
30B NVIDIA Nemotron 3.5 Lightning with Mixture‑of‑Experts
NVIDIA‑Nemotron‑3.5‑Lightning‑30B‑A3B is a 30 billion‑parameter LLM with 3 billion active parameters. It uses a hybrid nemotron_h_moe architecture with interleaved Mamba‑2 and MoE layers. The model is released in GGUF format for local inference.
Good at
- Large context handling up to 1 048 576 tokens
- Mixture‑of‑Experts for efficient inference
- High‑quality text generation
Limits
- Requires large VRAM for full precision
- No vision support
- License is OpenMDW‑1.1
Best for
- Customizing LLMs for specialized tasks
- Fine‑tuning with SFT or RLHF
- Deploying on high‑end GPUs
Who it’s for. Developers and researchers who need a large, flexible LLM for fine‑tuning or custom deployment. The model is suitable for commercial use under the OpenMDW‑1.1 license.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_M | 19.4 GB | 27.0 GB | Too big | Get file → | |
| IQ2_XXS | 19.4 GB | 27.0 GB | Too big | Get file → | |
| IQ2_M | 19.4 GB | 27.0 GB | Too big | Get file → | |
| IQ3_XXS | 19.8 GB | 27.4 GB | Too big | Get file → | |
| IQ3_S | 21.2 GB | 28.8 GB | Too big | Get file → | |
| IQ4_NL | 21.2 GB | 28.8 GB | Too big | Get file → | |
| Q3_K_XL | 21.2 GB | 28.8 GB | Too big | Get file → | |
| MXFP4_MOE | 23.2 GB | 30.8 GB | Too big | Get file → | |
| Q4_K_S | 24.5 GB | 32.1 GB | Too big | Get file → | |
| Q4_K_M | 25.3 GB | 32.9 GB | Too big | Get file → | |
| Q4_K_XL | 25.5 GB | 33.1 GB | Too big | Get file → | |
| Q5_K_S | 26.2 GB | 33.8 GB | Too big | Get file → | |
| Q5_K_M | 30.2 GB | 37.8 GB | Too big | Get file → | |
| Q5_K_XL | 30.4 GB | 38.0 GB | Too big | Get file → | |
| Q8_0 | 35.0 GB | 42.6 GB | Too big | Get file → | |
| Q6_K_XL | 35.0 GB | 42.6 GB | Too big | Get file → | |
| Q8_K_XL | 38.6 GB | 46.2 GB | Too big | Get file → | |
| BF16 | 65.9 GB (2 parts) | 73.5 GB | Too big | Get file → |
For 8 GB VRAM, use IQ1_M or IQ2_XXS (≈27 GB memory at 8k context). For 12 GB VRAM, IQ4_K_S or Q3_K_XL (≈32 GB). For 16 GB VRAM, Q4_K_M or Q5_K_S (≈33 GB). The estimates are approximate.
What GPU do you need for Nvidia-Nemotron-3.5-Lightning-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 MXFP4_MOE (23.2 GB file, about 30.8 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_K_XL (38.6 GB file, about 46.2 GB in use).
Which Nvidia-Nemotron-3.5-Lightning-30B-A3B quant should you choose?
Q4_K_M is the usual starting point for Nvidia-Nemotron-3.5-Lightning-30B-A3B: it balances file size and quality well. Here it is a 25.3 GB file.
The Q8_0 file is 35.0 GB, about 1.4 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is IQ1_M at 19.4 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about Nvidia-Nemotron-3.5-Lightning-30B-A3B
How much VRAM does Nvidia-Nemotron-3.5-Lightning-30B-A3B need?
About 32.9 GB at Q4_K_M with an 8K context window. That is the 25.3 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 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 27.0 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Nvidia-Nemotron-3.5-Lightning-30B-A3B GGUF?
It depends on the quant: IQ1_M is 19.4 GB, IQ2_XXS is 19.4 GB, IQ2_M is 19.4 GB, IQ3_XXS is 19.8 GB, IQ3_S is 21.2 GB, IQ4_NL is 21.2 GB and 12 more.
What context length does Nvidia-Nemotron-3.5-Lightning-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 7.1 GB at 8K and 28.5 GB at 32K.
Where can I download Nvidia-Nemotron-3.5-Lightning-30B-A3B GGUF?
The files are on Hugging Face in the unsloth/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-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 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?
Use IQ1_M or IQ2_XXS quantizations; they require about 27 GB of VRAM at 8k context, so 8 GB is insufficient.
Does the model support image input?
No, the model has no vision capabilities.
Is the model free to use?
The model is released under the OpenMDW‑1.1 license, which allows commercial use.