NVIDIA-Nemotron-3-Super-120B-A12B-BF16: hardware requirements
NVIDIA-Nemotron-3-Super-120B-A12B-BF16 needs about 90 GB of memory at Q4 quantization — a 86 GB download plus room for context and overhead. That fits a machine with 128GB RAM.
This is a mixture-of-experts model. Only a fraction of its 124B parameters are used for each token, so it runs faster than its size suggests — but all of it still has to fit in memory. The speed is that of a small model; the memory requirement is that of a large one. More on this →
Will it run on your machine?
Download sizes by quantization
Lower quantization means a smaller file and less memory, at some cost to quality. Q4_K_M is the usual starting point. What is this?
| Quant | Download | RAM @ 4K | RAM @ 32K | Source |
|---|---|---|---|---|
| Q4_K_Mrecommended | 86 GB | 90 GB | 92 GB | measured |
| Q5_K_M | 96 GB | 99 GB | 102 GB | measured |
| Q6_K | 113 GB | 117 GB | 119 GB | measured |
| Q8_0 | 131 GB | 135 GB | 138 GB | estimated |
| BF16 | 247 GB | 251 GB | 253 GB | estimated |
“Measured” sizes are read from published GGUF files (mradermacher/NVIDIA-Nemotron-3-Super-120B-A12B-BF16-i1-GGUF). “Estimated” sizes are derived from the parameter count and are typically within a few percent. The RAM columns add the context cache and runtime overhead to the download size.