NE

NVIDIA-Nemotron-3.5-Lightning-30B-A3B

NemotronGGUF by ggml-orgUpdated 2026-10-19
Download on Hugging Face
Parameters31.6B
Context1024K
Layers53
Architecturenemotron_h_moe
Smallest quant1.2 GB
Quick answer

NVIDIA-Nemotron-3.5-Lightning-30B-A3B needs about 11.7 GB of memory at BF16 (a 4.1 GB file) with an 8K context window, so it fits on a 12 GB GPU. This repository has 6 quants, from 1.2 GB (Q4_0) to 63.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 28.5 GB.

NVIDIA Nemotron 3.5 Lightning 30B A3B – 30B parameter LLM with MoE

NVIDIA Nemotron 3.5 Lightning 30B A3B is a 30 billion‑parameter language model using a hybrid Mamba‑2 and MoE architecture. It contains 3 billion active parameters and supports up to 1 048 576 context tokens. The model is released in GGUF format and is ready for commercial use.

Good at

  • Large context handling
  • Mixture‑of‑Experts efficiency
  • High‑capacity language modeling

Limits

  • Requires large VRAM for full precision
  • No vision or audio capabilities
  • License is OpenMDW‑1.1

Best for

  • Customizing domain‑specific agents
  • Fine‑tuning for specialized tasks
  • Deploying on high‑end GPUs

Who it’s for. Developers and researchers who need a high‑capacity LLM with a mixture‑of‑experts architecture for fine‑tuning or custom deployment. The model is suitable for commercial applications that require large context windows and high parameter counts.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_01.2 GB 8.8 GB
Too big Get file →
Q8_02.2 GB 9.8 GB
Too big Get file →
BF164.1 GB 11.7 GB
Too big Get file →
Q4_018.9 GB 26.5 GB
Too big Get file →
Q8_033.6 GB 41.2 GB
Too big Get file →
BF1663.2 GB 70.8 GB
Too big Get file →

For 8 GB VRAM, use the Q4_0 quantized file (1.16 GB, ~8.8 GB VRAM at 8 k context). For 16 GB VRAM, Q8_0 (2.18 GB, ~9.8 GB VRAM) or Q4_0 larger (18.9 GB, ~26.5 GB VRAM) can be used. BF16 variants require 11.7 GB or more VRAM and are suitable for larger GPUs.

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: the best quant that fits is BF16 (4.1 GB file, about 11.7 GB in use).
  • 16 GB VRAM: the best quant that fits is BF16 (4.1 GB file, about 11.7 GB in use).
  • 24 GB VRAM: the best quant that fits is BF16 (4.1 GB file, about 11.7 GB in use).
  • 32 GB VRAM: the best quant that fits is Q4_0 (18.9 GB file, about 26.5 GB in use).
  • 48 GB VRAM: the best quant that fits is Q8_0 (33.6 GB file, about 41.2 GB in use).

Which NVIDIA-Nemotron-3.5-Lightning-30B-A3B quant should you choose?

BF16 (4.1 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

If memory is tight, the smallest file is Q4_0 at 1.2 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturenemotron_h_moe
Parameters31.6B (approximate, mixture of experts)
Active parameters3B
Layers53
Trained context1,048,576 tokens
KV cache (fp16)848 KB per token

Source

Quantized byggml-org
FamilyNemotron
Licenseother
Downloads (HF)196,364
Last updated2026-10-19

Frequently asked questions about NVIDIA-Nemotron-3.5-Lightning-30B-A3B

How much VRAM does NVIDIA-Nemotron-3.5-Lightning-30B-A3B need?

About 11.7 GB at BF16 with an 8K context window. That is the 4.1 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 8.8 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: Q4_0 is 1.2 GB, Q8_0 is 2.2 GB, BF16 is 4.1 GB, Q4_0 is 18.9 GB, Q8_0 is 33.6 GB, BF16 is 63.2 GB.

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 ggml-org/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 NVIDIA‑Nemotron‑3.5‑Lightning‑30B‑A3B on 8 GB VRAM?

Use the Q4_0 quantized GGUF file (1.16 GB). It requires about 8.8 GB VRAM at 8 k context, which fits within 8 GB with some overhead.

Does this model support vision or audio inputs?

No, the model has no vision or audio capabilities. It is a text‑only language model.

What license governs the use of this model?

Use is governed by the OpenMDW‑1.1 model license, allowing commercial use under its terms.

Other GGUFs of NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16

Similar models

Browse all →

More from ggml-org

All by ggml-org →

More Nemotron models

All Nemotron →

Other medium models

All medium models →