Loading…
Loading…
Models / NVIDIA-Nemotron-3-Super
4 models (120B to 120B) · ↓ 1.8M downloads a month · from $1.10/h on 2x RTX 4090 right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalnvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | vLLM | 247 GB | 277.7 GB | $5.33/h 8x Nvidia A100 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalnvidia/NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 | vLLM | 80 GB | 92.3 GB | $1.10/h 2x RTX 4090 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalnvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8 | vLLM | 128 GB | 145.7 GB | $1.38/h 8x RTX 3090 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalnvidia/NVIDIA-Nemotron-3-Super-120B-A12B-Base-BF16 | vLLM | 247 GB | 277.7 GB | $5.33/h 8x Nvidia A100 |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").