NE

NVIDIA Nemotron Labs 3 Puzzle 75B A9B

NemotronGGUF by RemySkyeUpdated 2026-07-14
Download on Hugging Face
Parameters78.3B
Context1024K
Layers90
Architecturenemotron_h_moe
Smallest quant31.5 GB
Quick answer

NVIDIA Nemotron Labs 3 Puzzle 75B A9B needs about 64.2 GB of memory at Q4_K_M (a 51.6 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 16 quants, from 31.5 GB (Q2_K) to 156.8 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 48.3 GB.

NVIDIA Nemotron Labs 3 Puzzle 75B A9B – 78.3B parameter LLM with MoE

NVIDIA Nemotron Labs 3 Puzzle 75B A9B is a deployment‑optimized large language model with a hybrid MoE architecture. It supports reasoning and tool use, and offers a 1,048,576 token context window. The model is released under the OpenMDW‑1.1 license.

Good at

  • Reasoning tasks
  • Tool‑based interactions
  • Long‑context generation

Limits

  • No vision support
  • Requires specialized llama.cpp build
  • Large VRAM needed for higher quants

Best for

  • Interactive reasoning agents
  • Complex coding assistance
  • Multilingual long‑form content

Who it’s for. Developers building reasoning‑heavy applications that need a large context window and MoE efficiency. The model is suitable for commercial use under the OpenMDW‑1.1 license.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K31.5 GB 44.0 GB
Too big Get file →
Q3_K_S35.7 GB 48.2 GB
Too big Get file →
Q3_K_M40.6 GB 53.1 GB
Too big Get file →
IQ4_XS43.0 GB 55.5 GB
Too big Get file →
Q3_K_L44.2 GB 56.8 GB
Too big Get file →
Q4_044.5 GB 57.1 GB
Too big Get file →
Q4_K_S46.3 GB 58.9 GB
Too big Get file →
Q4_149.3 GB 61.9 GB
Too big Get file →
Q4_K_M51.6 GB 64.2 GB
Too big Get file →
Q5_054.2 GB 66.8 GB
Too big Get file →
Q5_K_S54.9 GB 67.5 GB
Too big Get file →
Q5_K_M58.7 GB 71.3 GB
Too big Get file →
Q5_159.1 GB 71.6 GB
Too big Get file →
Q6_K67.2 GB 79.8 GB
Too big Get file →
Q8_083.5 GB 96.0 GB
Too big Get file →
BF16156.8 GB 169.4 GB
Too big Get file →

For 8 GB VRAM, use Q4_0 (file size 44.49 GB, estimated 57.1 GB memory). For 12 GB VRAM, Q4_K_M (51.6 GB, 64.2 GB memory) is a good balance. For 24 GB VRAM, Q5_K_M (58.68 GB, 71.3 GB memory) provides higher quality.

What GPU do you need for NVIDIA Nemotron Labs 3 Puzzle 75B A9B?

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: no quant fits at an 8K context.
  • 48 GB VRAM: the best quant that fits is Q2_K (31.5 GB file, about 44.0 GB in use).

Which NVIDIA Nemotron Labs 3 Puzzle 75B A9B quant should you choose?

Q4_K_M is the usual starting point for NVIDIA Nemotron Labs 3 Puzzle 75B A9B: it balances file size and quality well. Here it is a 51.6 GB file.

The Q8_0 file is 83.5 GB, about 1.6 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 Q2_K at 31.5 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

Architecturenemotron_h_moe
Parameters78.3B (approximate, mixture of experts)
Active parameters9B
Layers90
Trained context1,048,576 tokens
KV cache (fp16)1440 KB per token

Source

Quantized byRemySkye
FamilyNemotron
Licenseother
Downloads (HF)738
Last updated2026-07-14

Frequently asked questions about NVIDIA Nemotron Labs 3 Puzzle 75B A9B

How much VRAM does NVIDIA Nemotron Labs 3 Puzzle 75B A9B need?

About 64.2 GB at Q4_K_M with an 8K context window. That is the 51.6 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 Labs 3 Puzzle 75B A9B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 44.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of NVIDIA Nemotron Labs 3 Puzzle 75B A9B GGUF?

It depends on the quant: Q2_K is 31.5 GB, Q3_K_S is 35.7 GB, Q3_K_M is 40.6 GB, IQ4_XS is 43.0 GB, Q3_K_L is 44.2 GB, Q4_0 is 44.5 GB and 10 more.

What context length does NVIDIA Nemotron Labs 3 Puzzle 75B A9B 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 12.1 GB at 8K and 48.3 GB at 32K.

Where can I download NVIDIA Nemotron Labs 3 Puzzle 75B A9B GGUF?

The files are on Hugging Face in the RemySkye/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-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 Labs 3 Puzzle 75B A9B 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 the Q4_0 quant; it requires about 57.1 GB memory at 8 k context, so 8 GB VRAM is insufficient. You need at least 12 GB VRAM for Q4_0.

Does this model support vision?

No, the model does not include vision capabilities.

What license applies to this model?

The model is governed by the OpenMDW‑1.1 license and is available for commercial use.

Other GGUFs of NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-BF16

Similar models

Browse all →

More from RemySkye

All by RemySkye →

More Nemotron models

All Nemotron →

Other large models

All large models →