How to list running Ollama models

The running model list shows which Ollama models are currently loaded into memory. It is different from the installed model list because unloaded models remain installed but do not consume runtime memory.

Use ollama ps for terminal checks and api/ps for automation. The output shows processor placement, context length, and unload time, which are the key signals for GPU and keep-alive debugging.

A blank process list is normal when no model is loaded. Run a short request first if the goal is to inspect a model's runtime placement.

1. Run a model so there is a process to inspect.

   $ ollama run gpt-oss:20b "Return only OK."
   OK

2. List loaded models with the CLI.

   $ ollama ps
   NAME           ID              SIZE     PROCESSOR    CONTEXT    UNTIL
   gpt-oss:20b    17052f91a42e    12 GB    100% GPU     131072     4 minutes from now

3. Read the same runtime list through the API.

   $ curl -s http://localhost:11434/api/ps
   {"models":[{"name":"gpt-oss:20b","processor":"100% GPU","context_length":131072}]}

4. Use the PROCESSOR column to verify GPU offload.

   $ printf '%s\n' '100% GPU'
   100% GPU

   Related: How to check Ollama GPU support

5. Stop the model when the process list should be empty.

   $ ollama stop gpt-oss:20b
   $ ollama ps
   NAME    ID    SIZE    PROCESSOR    CONTEXT    UNTIL