Loading…
Loading…
Models / NVIDIA-Nemotron-Nano
1 model (12B) · ↓ 1.9M downloads a month · from $0.14/h on RTX A4000 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 |
|---|---|---|---|---|
| NVIDIA Nemotron 3 Nano 4B BF16nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 | vLLM | 8.0 GB | 12.6 GB | $0.14/h RTX A4000 |
| NVIDIA Nemotron Nano 9B v2nvidia/NVIDIA-Nemotron-Nano-9B-v2 | vLLM | 18 GB | 24.1 GB | $0.34/h 2x RTX 5060 Ti |
| NVIDIA Nemotron Nano 9B v2 FP8nvidia/NVIDIA-Nemotron-Nano-9B-v2-FP8 | vLLM | 10 GB | 15.7 GB | $0.14/h RTX A4000 |
| NVIDIA Nemotron Nano 9B v2 NVFP4nvidia/NVIDIA-Nemotron-Nano-9B-v2-NVFP4 | vLLM | 7.8 GB | 13 GB | $0.14/h RTX A4000 |
| Instructnvidia/NVIDIA-Nemotron-Nano-12B-v2 | vLLM | 25 GB | 31.9 GB | $0.45/h Nvidia A40 |
| NVIDIA Nemotron Nano 9B v2 Japanesenvidia/NVIDIA-Nemotron-Nano-9B-v2-Japanese | vLLM | 18 GB | 24.1 GB | $0.34/h 2x RTX 5060 Ti |
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").