Ollama Cheat Sheet

Basic Commands

Command	Description
ollama serve	Start Ollama (if not running as a system service)
ollama run <model>	Run a model (and download it if it's not already installed)
ollama pull <model>	Pull/download a model without running it
ollama rm <model>	Remove/delete a model
ollama list	List all installed models
ollama ps	List models currently loaded in memory
ollama show <model>	Show details/information about a model
ollama cp <src> <dest>	Copy/rename a model

Gemma 2 Models (Best for Local Agents)

Model Size	Ollama Run Command	Notes
Gemma 2 2B	ollama run gemma2:2b	Perfect for background agents. Fast, lightweight (~1.5GB VRAM).
Gemma 2 9B	ollama run gemma2:9b	Excellent for RAG reasoning. High accuracy, requires ~6GB VRAM.
Gemma 2 27B	ollama run gemma2:27b	Heavy-duty reasoning. Requires ~18GB VRAM.

Interactive Prompt Commands

While inside ollama run <model>:

• /bye - Exit the prompt
• /? - Show help
• /set system <prompt> - Set a system prompt
• /set parameter temperature <value> - Adjust creativity (e.g., 0.7)

Creating Custom Models (Modelfile)

Create a file named Modelfile with:

FROM gemma4:e2b
PARAMETER temperature 0.7
SYSTEM You are a helpful programming assistant.

Then build it using: ollama create <custom-name> -f ./Modelfile