Inference
Running a trained model to produce an answer from the context it receives.
Illustrative conversation
- Did the model learn our coding standard after reading AGENTS.md once?
- It used the file during inference; its underlying parameters were not retrained by that request.
Definition
Inference is the act of using a trained model to make a prediction or generate a response from new input. During ordinary inference the model applies its learned parameters; it does not retrain them from the prompt. In an agent workflow, one inference request may return user-facing text, a structured tool call, or another intermediate decision.
Reviewed