Short answerRetrieval (RAG) looks up relevant passages from your documents at question time and gives them to the model. Fine-tuning changes the model itself using training examples. For a GPT that answers from documents, retrieval is usually the closer match and easier to update.
Retrieval (RAG)
Your files are split into passages. When someone asks a question, the system finds the most relevant passages and includes them in the prompt. Update a file and the answers can change the same day.
Fine-tuning
You train a model on example conversations so it behaves in a certain way. It teaches style and format well. It is a poor way to store facts that change.
| Retrieval | Fine-tuning | |
|---|---|---|
| Best for | Facts from documents | Consistent style or task behavior |
| Updating | Replace a file | Retrain |
| Citing sources | Easy | Hard |
| Effort | Lower | Higher |
For a GPT replacement
Put behavior in the instructions and facts in retrieved files. GPT Migrator hosting uses simple retrieval over your uploaded files. See moving a knowledge base.
Check your GPT before you move it
Paste your instructions, file names and Actions schema into the free checker. It runs in your browser and builds a portable bundle. Nothing is uploaded.
Open the free checker