Local LLM runtime
Ollama
Ollama is a MIT-licensed local-LLM server-capable runtime for Linux · macOS · Windows, running on CUDA · Metal · CPU · ROCm, with an OpenAI-compatible API. Ease of use 5/5 in our sourced ratings.
- License
- MIT
- Operating systems
- Linux · macOS · Windows
- Backends
- CUDA · Metal · CPU · ROCm
- Install
- binary
- Ease of use
- 5/5
- Latest version
- v0.31.2
- GitHub stars
- 175,773
- Engine
- llama.cpp
- Graphical app (GUI)
- Server mode
- OpenAI-compatible API
- CPU offload
- Multi-GPU
- Speculative decoding
- LoRA support
- KV-cache quantization
- One-line install
As of 2026-07-11 · source · GitHub
Quant-format support
What the feasibility gate checks before any VRAM math — per format and backend.
| Format | Backend | Supported |
|---|---|---|
| gguf | CPU | yes |
| gguf | CUDA | yes |
| gguf | Metal | yes |
| gguf | ROCm | yes |
| gguf | Vulkan | yes |
Software that runs on Ollama
Frontends, assistants and harnesses that can use Ollama as their local backend.
Best Ollama models by VRAM
Which models load cleanly at each memory budget, computed per tier.
Ollama compared
Side-by-side on install, backends, server mode and quant support.
Recommended in workflows
Will your machine run a model with Ollama?
Check with Ollama preselected →Support facts come from each runtime's own documentation, stamped and dated — how the math works.