NE

Nemotron H 47B Reasoning 128K

NemotronGGUF by bartowskiUpdated 2025-08-29
Download on Hugging Face
Parameters47B
Context1024K
Layers98
Architecturenemotron_h
Smallest quant0.0 GB
Quick answer

Nemotron H 47B Reasoning 128K needs about 55.0 GB of memory at Q4_K_M (a 28.2 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 28 quants, from 0.0 GB to 93.6 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 105.2 GB.

Nemotron H 47B Reasoning 128K – 47B parameter LLM with 128K context

Nemotron H 47B Reasoning 128K is a 47B parameter language model built on a Mamba‑2 and MLP hybrid architecture. It supports reasoning tasks and a 128K token context length. The model is available in multiple GGUF quantizations for llama.cpp.

Good at

  • Reasoning tasks
  • Long‑context handling up to 128K
  • Multilingual support (English, German, Spanish, French, Italian, Korean, Portuguese, Russian, Japanese, Chinese)

Limits

  • Requires large VRAM for full precision
  • No vision or audio input
  • License is NVIDIA Internal Scientific Research and Development Model License

Best for

  • Research prototypes needing deep reasoning
  • Applications requiring extended context
  • Developers building custom LLM pipelines

Who it’s for. Researchers and developers who need a large, reasoning‑capable model for text‑only tasks. The model is intended for research and development use only.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 26.8 GB
Too big Get file →
IQ2_XXS12.8 GB 39.7 GB
Too big Get file →
IQ2_XS14.1 GB 41.0 GB
Too big Get file →
IQ2_S14.6 GB 41.4 GB
Too big Get file →
IQ2_M15.9 GB 42.7 GB
Too big Get file →
Q2_K17.2 GB 44.0 GB
Too big Get file →
Q2_K_L18.2 GB 45.0 GB
Too big Get file →
IQ3_XXS18.3 GB 45.1 GB
Too big Get file →
IQ3_XS20.1 GB 46.9 GB
Too big Get file →
Q3_K_S20.6 GB 47.4 GB
Too big Get file →
IQ3_M21.0 GB 47.8 GB
Too big Get file →
Q3_K_M22.4 GB 49.2 GB
Too big Get file →
Q3_K_L23.8 GB 50.6 GB
Too big Get file →
Q3_K_XL24.7 GB 51.5 GB
Too big Get file →
IQ4_XS25.2 GB 52.0 GB
Too big Get file →
IQ4_NL26.6 GB 53.4 GB
Too big Get file →
Q4_026.8 GB 53.6 GB
Too big Get file →
Q4_K_S27.0 GB 53.8 GB
Too big Get file →
Q4_K_M28.2 GB 55.0 GB
Too big Get file →
Q4_K_L29.0 GB 55.8 GB
Too big Get file →
Q4_129.5 GB 56.3 GB
Too big Get file →
Q5_K_S32.3 GB 59.1 GB
Too big Get file →
Q5_K_M33.1 GB 60.0 GB
Too big Get file →
Q5_K_L33.8 GB 60.6 GB
Too big Get file →
Q6_K38.4 GB 65.2 GB
Too big Get file →
Q6_K_L38.9 GB 65.7 GB
Too big Get file →
Q8_049.7 GB 76.5 GB
Too big Get file →
BF1693.6 GB (3 parts) 120.4 GB
Too big Get file →

For 8‑GB VRAM, use IQ2_XXS (file size 12.85 GB, memory 39.7 GB at 8K context). For 16‑GB VRAM, use Q4_K_M (file size 28.19 GB, memory 55 GB). For 24‑GB VRAM, use Q4_K_L (file size 28.98 GB, memory 55.8 GB).

What GPU do you need for Nemotron H 47B Reasoning 128K?

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 (0.0 GB file, about 26.8 GB in use).
  • 48 GB VRAM: the best quant that fits is IQ3_M (21.0 GB file, about 47.8 GB in use).

Which Nemotron H 47B Reasoning 128K quant should you choose?

Q4_K_M is the usual starting point for Nemotron H 47B Reasoning 128K: it balances file size and quality well. Here it is a 28.2 GB file.

The Q8_0 file is 49.7 GB, about 1.8 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 the first one in the table at 0.0 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

Architecturenemotron_h
Parameters47B
Layers98
Trained context1,048,576 tokens
KV cache (fp16)3136 KB per token

Source

Quantized bybartowski
FamilyNemotron
Licenseother
Downloads (HF)2,497
Last updated2025-08-29

Frequently asked questions about Nemotron H 47B Reasoning 128K

How much VRAM does Nemotron H 47B Reasoning 128K need?

About 55.0 GB at Q4_K_M with an 8K context window. That is the 28.2 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 H 47B Reasoning 128K on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 26.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Nemotron H 47B Reasoning 128K GGUF?

It depends on the quant: file is 0.0 GB, IQ2_XXS is 12.8 GB, IQ2_XS is 14.1 GB, IQ2_S is 14.6 GB, IQ2_M is 15.9 GB, Q2_K is 17.2 GB and 22 more.

What context length does Nemotron H 47B Reasoning 128K 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 26.3 GB at 8K and 105.2 GB at 32K.

Where can I download Nemotron H 47B Reasoning 128K GGUF?

The files are on Hugging Face in the bartowski/nvidia_Nemotron-H-47B-Reasoning-128K-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 H 47B Reasoning 128K 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 H 47B Reasoning 128K on 8 GB VRAM?

No, the smallest quant requires 39.7 GB of VRAM at 8K context. You would need to use a smaller model or reduce context length.

Does this model support vision or audio input?

No, the model only accepts text input and does not have vision or audio capabilities.

What license applies to this model?

The model is governed by the NVIDIA Internal Scientific Research and Development Model License.

Similar models

Browse all →

More from bartowski

All by bartowski →

More Nemotron models

All Nemotron →

Other large models

All large models →