Nemotron Models — Hardware Requirements
45 Nemotron models from NVIDIA and the community, from the smallest that runs in 1.7 GB of VRAM at Q4 up to 560.5B parameters. Every row links to full quantization tables and GPU compatibility.
All Nemotron Models by Size
Community variants and fine-tunes(7)
| Model | Params | Runs from (Q4) | Context | Publisher | Quant downloads |
|---|---|---|---|---|---|
| Nemotron Content Safety Reasoning 4B | 4.3B | 3.6 GB | 131K | ||
| NVIDIA Nemotron 3.5 Bolt E4B BF16 | 8.0B | 5.6 GB | 66K | ||
| Nemotron Orchestrator 8B | 8.2B | 7.6 GB | 41K | ||
| Elbaz NVIDIA Nemotron 3 Nano 30B A3B PRISM | 30B | 23.5 GB | — | ||
| OpenReasoning Nemotron 32B | 32.8B | 24.3 GB | 131K | ||
| OpenCodeReasoning Nemotron 1.1 32B | 32.8B | 24.3 GB | 66K | ||
| NVIDIA Nemotron 3 Super 120B A12B BF16 Heretic | 120.7B | 72.9 GB | 262K |
How Nemotron Compares — llmrun Composite Score
NVIDIA Nemotron 3 Ultra 550B A55B BF16 is the highest-scoring Nemotron model with an llmrun Composite Score of 109.7 — #32 among 106 open models. The top proprietary model, Gemini 4 Carbon, scores 158.6. Click a model to see its full benchmark breakdown.
Gemini 4 Carbon · proprietary158.6
GPT 6 Astra · proprietary156.2
Claude Opus 5.5 · proprietary153.0
GPT 6.1 Sol · proprietary151.7
GPT 6.1 Terra · proprietary150.1
Kimi K3.1 Pro142.8
MiMo V2.6 Pro RL133.3
Kimi K3132.7
MiMo V2.6 Flash RL131.9
GLM 5.3130.1
DeepSeek V4.1 Flash129.4
DeepSeek V4 Pro 0813129.2
Frequently Asked Questions
- How much VRAM do I need to run a Nemotron model?
- The smallest Nemotron model, OpenMath Nemotron 1.5B, runs from 1.7 GB of VRAM at Q4_K_M (smaller low-bit quantizations start at 1.0 GB with a noticeable quality loss). Larger family members need proportionally more — see the table above for every model.
- Which Nemotron models can I run on a 16 GB GPU?
- 23 of 47 Nemotron models fit in 16 GB of VRAM at Q4_K_M, including NVIDIA Nemotron 3.5 Bolt E4B BF16, NVIDIA Nemotron Nano 9B v2, NVIDIA Nemotron 3 Nano 4B BF16.
- What is the most popular Nemotron model to run locally?
- NVIDIA Nemotron 3.5 Lightning 30B A3B BF16 is the most downloaded Nemotron model in local-friendly quantized formats. It runs from 19.3 GB of VRAM at Q4_K_M.
- How do Nemotron models score on benchmarks?
- NVIDIA Nemotron 3 Ultra 550B A55B BF16 leads the family with an llmrun Composite Score of 109.7, ranking #32 among 106 open models, while the top proprietary model, Gemini 4 Carbon, scores 158.6. See the comparison chart above for the full standings.