Updated: 28 September 2026 · Applies to: Ollama 0.34 on Ubuntu 24.04 / 26.04 LTS
Ollama is updated often, with support for new models and GPU fixes. On Linux you update it by running the install script again. This guide also shows how to remove Ollama completely.
Update Ollama
- Check your current version:
ollama -v
- Run the install script again. It replaces the program and keeps your models:
curl -fsSL https://ollama.com/install.sh | sh
- Restart the service and check the new version:
sudo systemctl restart ollama ollama -v
- Load a model to test:
ollama run llama3.2 "Say hello"
Your settings in /etc/systemd/system/ollama.service.d/override.conf (for example OLLAMA_HOST or OLLAMA_MODELS) stay in place. Check them with sudo systemctl cat ollama.
Update a model
Models are updated separately. Pull the model again to get its newest version:
ollama pull llama3.2
Uninstall Ollama
This deletes the program, the service and all downloaded models. Copy any custom model files you still need first.
- Stop and remove the service:
sudo systemctl stop ollama sudo systemctl disable ollama sudo rm /etc/systemd/system/ollama.service
- Remove the libraries (the path depends on where Ollama was installed,
/usr/local/lib,/usr/libor/lib):sudo rm -r $(which ollama | tr 'bin' 'lib')
- Remove the program:
sudo rm $(which ollama)
- Remove the models, the service user and its group:
sudo userdel ollama sudo groupdel ollama sudo rm -r /usr/share/ollama
If you moved your models with OLLAMA_MODELS (How to move Ollama models to another disk (OLLAMA_MODELS)?), also delete that folder if you no longer need it.
If the update fails
- The script cannot download: check that the server has internet access (
curl -I https://ollama.com) and DNS works. - Ollama runs on the CPU after the update: restart the service and check How to check that Ollama is using the GPU (and fix slow CPU-only answers)?.
- The service does not start: read
journalctl -e -u ollama.
Frequently asked questions
Will an update delete my models?
No. Models are kept in the models folder and are not touched by the install script.
How do I install a specific version?
Set OLLAMA_VERSION when you run the install script, for example curl -fsSL https://ollama.com/install.sh | OLLAMA_VERSION=0.5.7 sh. Version numbers are listed on Ollama's releases page.
Official documentation: Ollama documentation: Linux.
Want your own private AI without the setup work?
- Private LLM installation: we install and configure Ollama, the GPU driver and a chat interface on your own server.
- GPU dedicated servers: NVIDIA GPU servers for AI inference and training.
- AI implementation services: private LLMs, RAG, n8n automation and API integrations built on your own servers.
Prefer a hand with the setup? Our engineers can do it for you: Hire an Expert, or use our on-demand server management.