Nemotron 3 Super 120B-A12B
Nemotron 3 Super 120B-A12B loads in 86.8 GB at Q4_K_M and decodes at the speed of its 12B active path, comfortably from 128 GB of memory, Macs included, starting at the MacBook Pro M5 Max 128GB. Below is the memory math we ran for it, the cheapest card that fits, and the fastest machine we track.
How much memory does Nemotron 3 Super 120B-A12B need?
Nemotron 3 Super 120B-A12B is a 120B mixture-of-experts model (12B active per token) built for agentic workflows, long context, reasoning. The weights alone take 86.8 GB at Q4_K_M; add context and the OS, and it belongs in the 128 GB memory class. Every number below comes from the ModelFit engine — speeds are estimates, labeled est., because we do not benchmark hardware ourselves.
KV = fp16 estimate (q8_0 cache roughly halves it). "Comfortable" = weights + KV within the engine's tiered budget (~70-85% of memory).
Which machines run Nemotron 3 Super 120B-A12B?
Verdicts come from the same engine that powers our fit checker — memory capacity first, bandwidth for speed. How we compute these numbers · Hugging Face model card
More Nemotron models
Frequently asked questions
How much memory does Nemotron 3 Super 120B-A12B need?
86.8 GB for the Q4_K_M weights, plus ~6.0 GB of KV-cache at 16k context, about 92.8 GB total. Comfortable from 128 GB of VRAM or unified memory.
Does Nemotron 3 Super 120B-A12B run on a Mac?
Yes, from the MacBook Pro M5 Max 128GB (~18 tok/s est.). Unified memory means the RAM budget is the only limit.
What is the cheapest GPU for Nemotron 3 Super 120B-A12B?
The AMD Ryzen AI Max+ 395 (Strix Halo) is the cheapest tracked card that runs Nemotron 3 Super 120B-A12B comfortably, ~8 tok/s est. at ~$3,847 used (as of 2026-07-31).
Cite this page
ModelFit: Nemotron 3 Super 120B-A12B, specs, memory math and hardware verdicts. https://modelfit.io/models/nemotron-3-super-120b/ (dataset updated 2026-09-29, CC BY 4.0).