Guides · Updated 2026-10-17

How to install Ollama on Windows (2026)

Install Ollama on Windows 10 or 11, run your first local model, check it's using your GPU, and move models to another drive.

Ollama is the easiest way to run open AI models on a Windows PC. It installs as a normal app, runs in the background, and gives you the ollama command in PowerShell or Command Prompt.

What you need

• Windows 10 22H2 or newer (Home or Pro).
• NVIDIA: driver 551.61 or newer.
• AMD Radeon: a ROCm v7 / HIP-capable driver, or any Vulkan-capable Radeon driver (Vulkan is on by default and is the fallback for cards like the RX 6000 series).
• Disk space: about 4 GB for Ollama itself, plus 2-25 GB per model.

No admin rights are needed: Ollama installs into your user folder.

Step 1: Download and install

1. Go to ollama.com/download and download OllamaSetup.exe.
2. Run it and click Install. When it finishes, Ollama starts in the background and appears in the system tray.

Step 2: Run your first model

Open PowerShell and run:

ollama run gemma4:e4b

The first run downloads the model (about 6.6 GB), then drops you into a chat. Type a question, and type /bye to exit. Not sure which model suits your card? Use the GPU picker.

Step 3: Check it’s using your GPU

While a model is loaded, run:

ollama ps

The PROCESSOR column should say 100% GPU. If it shows something like 48%/52% CPU/GPU, the model is too big for your VRAM and part of it is running from system RAM, which is much slower. Pick a smaller model or a smaller tag.

Step 4 (optional): Store models on another drive

Models live in %HOMEPATH%\.ollama by default. To move them:

1. Quit Ollama from the tray icon.
2. Search Windows for environment variables and choose Edit environment variables for your account.
3. Add a variable named OLLAMA_MODELS with a path such as D:\ollama-models.
4. Start Ollama again from the Start menu.

You can also install the app itself elsewhere: OllamaSetup.exe /DIR="D:\Ollama".

Use it from other apps

Ollama serves an API at http://localhost:11434, which tools like Open WebUI, Continue, and many others connect to. Quick test from PowerShell:

(Invoke-WebRequest -method POST -Body '{"model":"gemma4:e4b","prompt":"Say hi","stream":false}' -uri http://localhost:11434/api/generate).Content | ConvertFrom-Json

Troubleshooting

• Logs: press Win+R and open %LOCALAPPDATA%\Ollama. server.log shows GPU detection.
• Boxes instead of progress bars: an older Windows 10 terminal font; switch font or use Windows Terminal.
• AMD mixed iGPU + dGPU picks the wrong GPU: set GGML_VK_VISIBLE_DEVICES to your discrete GPU’s index.

Source: Ollama Windows docs.