Local AI Engineering with Ollama

Listing Saved and Loaded Models

Once you start pulling models, you'll accumulate them fast. Each one sits on disk taking gigabytes, and any one of them can be loaded into memory at a moment's notice. You need to know how to inspect what you have, free up memory, and clean up disk space.

We've already seen this before. ollama list shows every model on disk:

ollama list

Output example:

NAME              ID              SIZE      MODIFIED       
granite3.3:2b     07bd1f170855    1.5 GB    2 hours ago
llama3.2:3b       a80c4f17acd5    2.0 GB    3 days ago
qwen2.5:7b        845dbda0ea48    4.7 GB    1 week ago

(i) Reminder: SIZE here is disk size, not memory size. The same model loaded into RAM is bigger because of the KV cache. You can see this after running ollama ps: