NVIDIA · llama
Can I run nemotron?
Llama-3.1-Nemotron-70B-Instruct is a large language model customized by NVIDIA to improve the helpfulness of LLM generated responses to user queries.
Parameter counts and memory belong to each build below. A family can contain several sizes and tasks; fitting in memory is not a quality ranking.
Calculate memory for a build and contextnemotron system requirements
Memory needed at 4K context, counting model weights, KV cache, compute buffers and runtime overhead. Download sizes come from the official Ollama registry.
| Build | Quantization | Download | Memory needed |
|---|---|---|---|
| nemotron:70b-instruct-q4_070B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q4_0 | 37.2 GB | ~41.8 GB |
| nemotron:70b70B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q4_K_M | 39.6 GB | ~44.4 GB |
| nemotron:70b-instruct-q5_K_M70B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q5_K_M | 46.5 GB | ~51.9 GB |
| nemotron:70b-instruct-q6_K70B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q6_K | 53.9 GB | ~59.9 GB |
| nemotron:70b-instruct-q8_070B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q8_0 | 69.8 GB | ~77.1 GB |
| nemotron:70b-instruct-fp1670B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | F16 | 131.4 GB | ~143.6 GB |
Showing the 6 smallest non-alias builds. Choose any build and change the context. A context above the build's documented maximum is a hypothetical calculation, not a supported setting.
How to run nemotron locally
- 1.Install Ollama for Windows, macOS or Linux.
- 2.Download this exact build in a terminal:
ollama pull nemotron:70b-instruct-q4_0 - 3.Check the source for the build's task. For text generation, run
ollama run nemotron:70b-instruct-q4_0. Base/text builds may need a prompt template; an instruct build is usually the appropriate starting point for a conversation.
Frequently asked questions about nemotron
- How much VRAM do I need to run nemotron?
- At 4K context, the smallest published build of nemotron (Q4_0) needs roughly 41.8 GB of GPU memory once weights, KV cache, compute buffers and runtime overhead are counted. Bigger quantizations and longer context windows need more. It can also run partly or entirely on the CPU using system RAM, more slowly.
- Can I run nemotron without a dedicated GPU?
- A supported CPU runtime can use system RAM. The smallest listed build needs an estimated 41.8 GB at 4K context, plus headroom for your operating system. Check the build's runtime requirements. We have not measured its speed on your CPU.
- How big is the nemotron download?
- The smallest published build is 37.2 GB. There are 6 builds in total, the largest being 131.4 GB. Leave some extra free disk space beyond the download itself.
- How do I run nemotron locally?
- Install Ollama, then download a specific build with "ollama pull nemotron:70b-instruct-q4_0". For a build that supports text generation, use ollama run with the same tag. Check the model source for its task and prompt format.
Will nemotron run on your PC?
Scan your hardware once and get a verdict for this model and every build — free, no account.
Compare with similar models
Or start from your memory: 8 GB VRAM, 12 GB VRAM, 16 GB VRAM, 24 GB VRAM or a 16 GB Mac.
Model data from the official Ollama registry · verified 13/10/2026