🧠 Prompt vs Completion — AI Input/Output Terminology by Context

Standard Terms in LLMs

Input Term	Output Term	Used In
Prompt	Completion	GPT models, OpenAI API, text-based LLMs
Prompt	Response	Chat interfaces like ChatGPT
Prompt	Generation	NLP/text generation tasks (e.g. Hugging Face)
Prompt	Prediction	Instruct-tuned or few-shot models

Input Term	Output Term	Used When
Input	Inference Result	Generic ML models (e.g. image → label)
Input Features	Predictions	Supervised learning / regression
Seed Text / Image	Generation	Creative AI (text/image/music)
User Message	Assistant Reply	Chatbot interfaces
Input Text/Token	Output Text/Token	Token-level modeling (e.g. Transformers)
Query	Response	Search engines, retrieval-augmented generation

Completion is a core term in OpenAI's API and early GPT framing.
Inference result generalizes across all types of ML models.
Generation is preferred in creative or unsupervised contexts.
Terminology varies depending on whether the system is trained for prediction, classification, generation, or dialogue.