Model weights occupy disk space before Ollama can load them into memory. A public model pull downloads the chosen model's manifest and layers into the Ollama server's local model store.
A separate pull lets the download finish before a prompt is sent. The CLI can also download a missing model during ollama run, so models do not always need a separate pull before use.
Downloads consume disk space, and working memory is a separate requirement. A model can fit on disk yet fail to load into memory; a short response after the pull confirms that the downloaded model can process a prompt.
Steps to pull an Ollama model:
- Check the models already installed on the Ollama server.
$ ollama list NAME ID SIZE MODIFIED
An empty list contains only the column headings. If the chosen name and tag already appear, pulling again checks for updates and reuses layers already in the store. A connection error means the Ollama server is unavailable.
Related: How to start the Ollama server - Pull the public qwen2.5:0.5b model into the store.
$ ollama pull qwen2.5:0.5b pulling manifest ##### snipped ##### verifying sha256 digest writing manifest successThe 0.5b tag selects the small Qwen2.5 text model, with a download of about 398 MB. Larger models can require many gigabytes of free disk space. Model files are stored on the server the CLI connects to, including when that server is remote.
Related: Qwen2.5 model tag - Confirm that qwen2.5:0.5b appears in the installed model list.
$ ollama list NAME ID SIZE MODIFIED qwen2.5:0.5b a8b0c5157701 397 MB Less than a second agoThe NAME includes the tag; its row confirms the pull has registered the model in this server's store. The MODIFIED value changes with time.
- Run the downloaded model with a short arithmetic prompt.
$ ollama run qwen2.5:0.5b "What is 17 plus 25? Answer in one short sentence." 17 plus 25 equals 42.
The answer should express that the sum is 42; wording can vary. A returned answer confirms that the downloaded model can load and process the prompt. The model files remain in the store for later use.
Related: How to run a model in Ollama