IBM · granite
Can I run granite4?
Granite 4 features improved instruction following (IF) and tool-calling capabilities, making them more effective in enterprise applications.
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 contextgranite4 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 |
|---|---|---|---|
| granite4:350m-h-q8_0350M parameters (tag)1024K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q8_0 | 0.3 GB | ~1.2 GB |
| granite4:350m350M parameters (tag)32K documented maximum contextBuild sourceCatalog checked 2026-10-13 | BF16 | 0.7 GB | ~1.6 GB |
| granite4:1b-h-q8_01B parameters (tag)1024K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q8_0 | 1.5 GB | ~2.9 GB |
| granite4:3b3B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | Q4_K_M | 2.0 GB | ~4.1 GB |
| granite4:1b1B parameters (tag)128K documented maximum contextBuild sourceCatalog checked 2026-10-13 | BF16 | 3.0 GB | ~4.9 GB |
Showing the 5 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 granite4 locally
- 1.Install Ollama for Windows, macOS or Linux.
- 2.Download this exact build in a terminal:
ollama pull granite4:350m-h-q8_0 - 3.Check the source for the build's task. For text generation, run
ollama run granite4:350m-h-q8_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 granite4
- How much VRAM do I need to run granite4?
- At 4K context, the smallest published build of granite4 (Q8_0) needs roughly 1.2 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 granite4 without a dedicated GPU?
- A supported CPU runtime can use system RAM. The smallest listed build needs an estimated 1.2 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 granite4 download?
- The smallest published build is 0.3 GB. There are 5 builds in total, the largest being 3.0 GB. Leave some extra free disk space beyond the download itself.
- How do I run granite4 locally?
- Install Ollama, then download a specific build with "ollama pull granite4:350m-h-q8_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 granite4 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