Reference · Glossary

Completion

Last updated

The model's **generated continuation** of a prompt — the next tokens it predicts, whether a single phrase or a full reply.

When to use

Explaining API response fields, base-model behavior, or why the model "autocompletes" rather than truly "understands."

When not to

Chat products wrap completions in roles (system/user/assistant) — the underlying mechanism is still next-token prediction.

Example

Prompt: "The capital of France is" → completion: " Paris." Same engine powers long assistant answers.