NE

Nemotron H 8B Reasoning 128K

NemotronGGUF by bartowskiUpdated 2025-08-29
Download on Hugging Face
Parameters8.1B
Context1024K
Layers52
Architecturenemotron_h
Smallest quant0.0 GB
Quick answer

Nemotron H 8B Reasoning 128K needs about 12.5 GB of memory at Q4_K_M (a 5.0 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 25 quants, from 0.0 GB to 16.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 27.9 GB.

Nemotron H 8B Reasoning 128K – large‑scale reasoning LLM

Nemotron H 8B Reasoning 128K is an 8.1B parameter model built on a Mamba‑2 and MLP hybrid architecture. It supports up to 128K context tokens and can generate reasoning traces before answering. The model is available in multiple GGUF quantizations.

Good at

  • Reasoning tasks
  • Long‑context text generation
  • Multilingual support

Limits

  • Requires large VRAM for high‑quality quants
  • No vision or audio input
  • License is NVIDIA Internal Scientific Research and Development

Best for

  • Research prototypes needing reasoning
  • Long‑form content creation
  • Academic experimentation

Who it’s for. Developers and researchers building or testing large language models. The model is intended for research and development only and is governed by NVIDIA’s internal license.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 7.5 GB
Tight Get file →
IQ2_M2.9 GB 10.4 GB
Too big Get file →
Q2_K3.2 GB 10.6 GB
Too big Get file →
IQ3_XXS3.3 GB 10.8 GB
Too big Get file →
IQ3_XS3.6 GB 11.1 GB
Too big Get file →
Q2_K_L3.7 GB 11.2 GB
Too big Get file →
Q3_K_S3.7 GB 11.2 GB
Too big Get file →
IQ3_M3.8 GB 11.3 GB
Too big Get file →
Q3_K_M4.0 GB 11.5 GB
Too big Get file →
Q3_K_L4.3 GB 11.8 GB
Too big Get file →
IQ4_XS4.5 GB 12.0 GB
Too big Get file →
IQ4_NL4.7 GB 12.2 GB
Too big Get file →
Q3_K_XL4.7 GB 12.2 GB
Too big Get file →
Q4_04.7 GB 12.2 GB
Too big Get file →
Q4_K_S4.8 GB 12.3 GB
Too big Get file →
Q4_K_M5.0 GB 12.5 GB
Too big Get file →
Q4_15.2 GB 12.7 GB
Too big Get file →
Q4_K_L5.4 GB 12.9 GB
Too big Get file →
Q5_K_S5.7 GB 13.1 GB
Too big Get file →
Q5_K_M5.8 GB 13.3 GB
Too big Get file →
Q5_K_L6.1 GB 13.6 GB
Too big Get file →
Q6_K6.7 GB 14.1 GB
Too big Get file →
Q6_K_L6.9 GB 14.4 GB
Too big Get file →
Q8_08.6 GB 16.1 GB
Too big Get file →
BF1616.2 GB 23.7 GB
Too big Get file →

For 8‑GB VRAM, use IQ2_M (2.93 GB) or Q2_K (3.16 GB). For 12‑GB VRAM, Q4_K_M (4.98 GB) or Q5_K_S (5.65 GB) are suitable. Higher‑quality quants such as Q6_K (6.66 GB) or BF16 (16.21 GB) require more memory.

What GPU do you need for Nemotron H 8B 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: the best quant that fits is the main file (0.0 GB file, about 7.5 GB in use).
  • 12 GB VRAM: the best quant that fits is IQ4_XS (4.5 GB file, about 12.0 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K_L (6.9 GB file, about 14.4 GB in use).
  • 24 GB VRAM: the best quant that fits is BF16 (16.2 GB file, about 23.7 GB in use).

Which Nemotron H 8B Reasoning 128K quant should you choose?

Q4_K_M is the usual starting point for Nemotron H 8B Reasoning 128K: it balances file size and quality well. Here it is a 5.0 GB file.

The Q8_0 file is 8.6 GB, about 1.7 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
Parameters8.1B
Layers52
Trained context1,048,576 tokens
KV cache (fp16)832 KB per token

Source

Quantized bybartowski
FamilyNemotron
Licenseother
Downloads (HF)1,806
Last updated2025-08-29

Frequently asked questions about Nemotron H 8B Reasoning 128K

How much VRAM does Nemotron H 8B Reasoning 128K need?

About 12.5 GB at Q4_K_M with an 8K context window. That is the 5.0 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 H 8B Reasoning 128K on an 8 GB GPU?

Yes, with the smaller files: the 0.0 GB file needs about 7.5 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Nemotron H 8B Reasoning 128K GGUF?

It depends on the quant: file is 0.0 GB, IQ2_M is 2.9 GB, Q2_K is 3.2 GB, IQ3_XXS is 3.3 GB, IQ3_XS is 3.6 GB, Q2_K_L is 3.7 GB and 19 more.

What context length does Nemotron H 8B 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 7.0 GB at 8K and 27.9 GB at 32K.

Where can I download Nemotron H 8B Reasoning 128K GGUF?

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

Yes, using the IQ2_M or Q2_K quantizations, which are 2.93 GB and 3.16 GB respectively. The estimated memory at 8K context is 10.4 GB and 10.6 GB, which is within 8 GB VRAM limits with efficient batching.

Does this model support vision or audio input?

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

What license applies to this model?

The model is governed by NVIDIA Internal Scientific Research and Development Model License, and is for research and development only.

Similar models

Browse all →

More from bartowski

All by bartowski →

More Nemotron models

All Nemotron →

Other small models

All small models →