Nemotron-3 Super (120B-A12B, Q4)
NVIDIA's flagship Nemotron-3 Super at 120B parameters, tuned for reasoning. Native tool calling, 128K context. Frontier-tier on-device reasoning for workstations and Mac Studio-class hardware — needs ~100GB of usable memory. No Apple Silicon build yet.
| At a glance | |
|---|---|
| Made by | NVIDIA |
| Download size | 86.1 GB |
| Parameters | 120B |
| Quantization | Q4_K_M |
| Context window | 131K tokens |
| Uses tools | Yes |
| Best for | Reasoning |
| Runs on | llama.cpp |
| License | NVIDIA Open Model |
| Added to the catalog | 2026-08-15 |
Gezel downloads this model the first time a gezel needs it, checks the file against its published hash, and applies tuned settings for it.