How to update an existing model to its latest version

What this does

Triggers a re-download of the designated model, fetching any newer manifests or binary blobs from the remote library. The local copy is replaced with the current release, eliminating stale weights or outdated configuration.

Steps

1. Confirm the currently installed version. Obtains the digest of the model already stored locally.

   ollama show llama3.2 | head -5
   

   Expected output: A manifest block displaying architecture, parameter count, and digest hash.

2. Pull the model again with the same name. Ollama resolves the latest available version and downloads only changed layers.

   ollama pull llama3.2
   

   Expected output: Resuming or downloading progress bars followed by success.

3. Compare the new digest to the old one. Confirms that the binary changed after the pull operation.

   ollama show llama3.2 | grep digest
   

   Expected output: A different hex string than the original value from step 1, or identical if already current.

• Record the local run evidence. Save the exact command, runtime or package version, model name if applicable, and observed output so the result can be reproduced later.

Verification

ollama list
# Expected: the model row displays a refreshed MODIFIED timestamp matching today's date or the pull time

Common failures

• already best - Model is fully current; Ollama reports no new layers to download, which is normal.
• network timeout - Slow or interrupted internet; retry the pull command.
• incomplete download - Disk write failure midway; re-run the pull to resume from the last checkpoint.
• model not found - Typo in name; run ollama list to retrieve the exact spelling.
• storage full during pull - Insufficient disk space for the update; free space and re-run.

Related guides

• How to pull and run your first open-source LLM using Ollama
• How to verify model integrity after downloading