NVIDIA-Nemotron-3-Ultra-550B-A55B needs about 409.1 GB of memory at Q4_K_M (a 379.6 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 5 quants, from 14.2 GB (Q8_0) to 436.2 GB (Q5_K_M). Longer context needs more: at 32K tokens the KV cache alone adds about 116.0 GB.
NVIDIA-Nemotron-3-Ultra-550B-A55B is a 549B-parameter Nemotron model in GGUF format, quantized by vultr. Its trained context length is 1024K tokens. The repository has 5 quant files, from 14.2 GB (Q8_0) to 436.2 GB (Q5_K_M).
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q8_0 | 14.2 GB | 43.7 GB | Too big | Get file → | |
| IQ3_S | 260.8 GB (2 parts) | 290.3 GB | Too big | Get file → | |
| IQ3_M | 265.7 GB (2 parts) | 295.2 GB | Too big | Get file → | |
| Q4_K_M | 379.6 GB (3 parts) | 409.1 GB | Too big | Get file → | |
| Q5_K_M | 436.2 GB (3 parts) | 465.7 GB | Too big | Get file → |
What GPU do you need for NVIDIA-Nemotron-3-Ultra-550B-A55B?
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 Q8_0 (14.2 GB file, about 43.7 GB in use).
Which NVIDIA-Nemotron-3-Ultra-550B-A55B quant should you choose?
Q4_K_M is the usual starting point for NVIDIA-Nemotron-3-Ultra-550B-A55B: it balances file size and quality well. Here it is a 379.6 GB file.
The Q8_0 file is 14.2 GB, about 0.0 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 Q8_0 at 14.2 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Frequently asked questions about NVIDIA-Nemotron-3-Ultra-550B-A55B
How much VRAM does NVIDIA-Nemotron-3-Ultra-550B-A55B need?
About 409.1 GB at Q4_K_M with an 8K context window. That is the 379.6 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-3-Ultra-550B-A55B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 43.7 GB, so you would need a larger GPU or CPU offloading.
What is the file size of NVIDIA-Nemotron-3-Ultra-550B-A55B GGUF?
It depends on the quant: Q8_0 is 14.2 GB, IQ3_S is 260.8 GB, IQ3_M is 265.7 GB, Q4_K_M is 379.6 GB, Q5_K_M is 436.2 GB.
What context length does NVIDIA-Nemotron-3-Ultra-550B-A55B 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 29.0 GB at 8K and 116.0 GB at 32K.
Where can I download NVIDIA-Nemotron-3-Ultra-550B-A55B GGUF?
The files are on Hugging Face in the vultr/NVIDIA-Nemotron-3-Ultra-550B-A55B-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-Ultra-550B-A55B 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.