Ollama Commands Cheat Sheet: Complete CLI Reference (2026)

Download a model without running it

ollama pull qwen3:8b

List all installed models (name, size, modified date)

ollama list

Remove a model to free disk space

ollama rm qwen3:14b

Show model details, parameters, and Modelfile

ollama show qwen3:8b

Show template and system prompt for a model

ollama show qwen3:8b --modelfile

Copy a model under a new local name

ollama cp qwen3:8b my-qwen

Check which models are loaded in VRAM right now

ollama ps

Unload a model from VRAM immediately

ollama stop qwen3:8b

/api/generate — single-turn completion

curl http://localhost:11434/api/generate \
  -d '{"model":"qwen3:8b","prompt":"What is the capital of France?","stream":false}'

/api/chat — multi-turn conversation with messages array

curl http://localhost:11434/api/chat \
  -d '{
    "model": "qwen3:8b",
    "messages": [
      {"role": "user", "content": "Why is the sky blue?"}
    ],
    "stream": false
  }'

OpenAI-compatible endpoint — drop-in replacement for OpenAI API

curl http://localhost:11434/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3:8b",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Python using the OpenAI library pointed at Ollama

from openai import OpenAI

client = OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")
response = client.chat.completions.create(
    model="qwen3:8b",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

List available models via API

curl http://localhost:11434/api/tags

Stream responses token-by-token (stream defaults to true)

curl http://localhost:11434/api/generate -d '{"model":"qwen3:8b","prompt":"Tell me a joke"}'

Example Modelfile — coding assistant with 8K context

FROM qwen3:8b

SYSTEM """
You are an expert software engineer. Write clean, well-commented code.
Always explain what the code does and why you chose this approach.
"""

PARAMETER temperature 0.3
PARAMETER num_ctx 8192
PARAMETER num_predict 2048

Create and run the custom model

ollama create coding-assistant -f Modelfile

ollama run coding-assistant

Edit the Ollama systemd service override

sudo systemctl edit ollama

Benchmark token speed for a model

ollama run qwen3:8b --verbose "write a 200 word essay about AI" 2>&1 | grep "eval rate"

Ask a one-shot question and get a plain text answer

ollama run qwen3:8b "what is 17 * 43?" --nowordwrap

Summarize a local file

cat report.txt | ollama run qwen3:8b "summarize this in 3 bullet points"

Translate a file to Spanish

cat document.txt | ollama run qwen3:8b "translate this to Spanish"

Check if Ollama API is up

curl -s http://localhost:11434/ | grep -i ollama

Pull multiple models in sequence

for m in qwen3:8b phi4:14b gemma3:12b; do ollama pull $m; done

Remove all models at once (careful: deletes everything)

ollama list | awk 'NR>1 {print $1}' | xargs -I{} ollama rm {}

Keep Ollama model loaded indefinitely (no idle unload)

OLLAMA_KEEP_ALIVE=0 ollama serve

Force a model to use GPU layers (in Modelfile)

PARAMETER num_gpu 99

Run Ollama with a custom host and port

OLLAMA_HOST=0.0.0.0:11435 ollama serve

Check if Ollama service is running

systemctl status ollama

View Ollama logs (Linux)

journalctl -u ollama -n 50 --no-pager

View Ollama logs (macOS)

cat ~/.ollama/logs/server.log | tail -50

Restart Ollama service (Linux)

sudo systemctl restart ollama

Check NVIDIA GPU and VRAM

nvidia-smi

Check AMD GPU and ROCm

rocm-smi

Confirm GPU is used (check GPU% column)

ollama ps

Check Ollama version

ollama --version

Run model with verbose token speed output

ollama run qwen3:8b --verbose