Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4: hardware requirements

Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 needs about 27 GB of memory at Q4 quantization — a 22 GB download plus room for context and overhead. That fits a machine with 64GB RAM.

Parameters
18B
Tier
Needs a good desktop
Licence
See model card
Released
2026-04-28

This is a mixture-of-experts model. Only a fraction of its 18B 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?

QuantDownloadRAM @ 4KRAM @ 32KSource
Published22 GB27 GB27 GBmeasured

“Measured” sizes are read from published GGUF files (nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4). “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.

Model card on HuggingFace →API specs & pricing →Verified 2026-08-09Tracked for 12 months after release