[ GLOSSARY ]

What is a large language model?

Facts as of Oct 18, 2026Last reviewed Oct 19, 20261 min read

Definition

AI concepts

Large language model

A large language model (LLM) is an AI model trained on large amounts of text to predict and generate language. It is the engine behind chatbots such as ChatGPT, Claude and Gemini.

Why it matters for schools

LLMs can explain, draft and solve, but they make mistakes and can state false things confidently. Quality varies by model: in one study of middle-school math explanations, GPT-4's were valid 94% of the time, against 10-43% for GPT-3.5. Solving is not teaching, either: in one tutoring benchmark, teaching instructions changed an AI's scaffolding more than raw solving strength did.

[ FAQ ]

Questions and answers

What is an LLM in simple terms?

A model trained on large amounts of text that predicts and generates language, used in chatbots.

Are all LLMs equally accurate?

No. In one study, GPT-4's math explanations were valid 94% of the time, against 10-43% for GPT-3.5.

Can students use any LLM at school?

Provider terms limit minors' use: Google's Gemini API terms bar services likely used by under-18s, and consumer Claude requires users to be 18 or older.

[ SOURCES ]