Meta AI · llama4
Can I run llama4?
Meta's latest collection of multimodal models.
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 contextllama4 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 |
|---|---|---|---|
| llama4:17b-scout-16e-instruct-q4_K_M17B parameters (tag)10240K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q4_K_M | 62.8 GB | ~69.0 GB |
| llama4:17b-scout-16e-instruct-q8_017B parameters (tag)10240K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q8_0 | 108.5 GB | ~118.3 GB |
| llama4:17b-scout-16e-instruct-fp1617B parameters (tag)10240K documented maximum contextBuild sourceCatalog checked 2026-10-13 | F16 | 202.4 GB | ~219.7 GB |
| llama4:17b-maverick-128e-instruct-q4_K_M17B parameters (tag)1024K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q4_K_M | 228.0 GB | ~247.4 GB |
| llama4:17b-maverick-128e-instruct-q8_017B parameters (tag)1024K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q8_0 | 398.4 GB | ~431.4 GB |
| llama4:17b-maverick-128e-instruct-fp1617B parameters (tag)1024K documented maximum contextBuild sourceCatalog checked 2026-10-13 | F16 | 748.1 GB | ~809.0 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 llama4 locally
- 1.Install Ollama for Windows, macOS or Linux.
- 2.Download this exact build in a terminal:
ollama pull llama4:17b-scout-16e-instruct-q4_K_M - 3.Check the source for the build's task. For text generation, run
ollama run llama4:17b-scout-16e-instruct-q4_K_M. Base/text builds may need a prompt template; an instruct build is usually the appropriate starting point for a conversation.
Frequently asked questions about llama4
- How much VRAM do I need to run llama4?
- At 4K context, the smallest published build of llama4 (Q4_K_M) needs roughly 69.0 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 llama4 without a dedicated GPU?
- A supported CPU runtime can use system RAM. The smallest listed build needs an estimated 69.0 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 llama4 download?
- The smallest published build is 62.8 GB. There are 6 builds in total, the largest being 748.1 GB. Leave some extra free disk space beyond the download itself.
- How do I run llama4 locally?
- Install Ollama, then download a specific build with "ollama pull llama4:17b-scout-16e-instruct-q4_K_M". 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 llama4 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