Download Ollama for Windows

Install the official Ollama app and run open‑source large language models (Llama 3, Mistral, Qwen 2.5, Gemma 2, Phi‑4) locally on your Windows PC — private, fast, and offline.

⬇️ Download Ollama for Windows (Official)

Official installer • Loading… • 64‑bit Windows

Why run LLMs locally with Ollama?

Keep your data on your machine while getting fast, reliable AI. Ollama makes local models simple for developers and power users alike.

Easy Windows setup
One‑command model downloads
Privacy‑first, offline usage
Open‑source and community‑driven

Quick start

After installing, open Command Prompt or PowerShell and pull your first model:

ollama pull llama3

Then run it locally:

ollama run llama3

Developers: head over to the Python section for API usage.

Minimum system requirements

OS	Windows 10/11 (64‑bit)
CPU	Modern multi‑core (Intel i5/AMD Ryzen 5 or better)
Memory	8 GB for smaller models; 16 GB+ recommended for 8B models
GPU	NVIDIA CUDA for best acceleration; DirectML support varies by model/driver
Disk space	~5–15 GB per model depending on quantization (e.g., 8B Q4 ≈ ~5–7 GB)
License	Open‑source (see upstream project licenses)

Model sizes and performance vary by parameter count and quantization; GPU VRAM (8 GB+) is recommended for speed, but CPU‑only works for smaller models.