What is an LLM (Large Language Model)?

Definition

An LLM (Large Language Model) is an artificial intelligence model trained on large amounts of text to generate, understand, and synthesize natural language. ChatGPT, Claude, Gemini, and Mistral are LLMs. They form the foundation of AI answer engines.

LLMs (Large Language Models) are neural networks trained at scale on text corpora from the web, books, and scientific articles. They learn to predict the next token in a sequence, enabling them to generate coherent, context-aware responses.

LLMs and search

Several search engines now integrate LLMs into their interface: Google with AI Overviews (powered by Gemini), Bing with Copilot (GPT-4), Perplexity with its own system. These engines generate a synthetic response rather than simply displaying a list of links.

What this means for brands

LLMs build their responses from their training corpus and, for the most recent models, from real-time web searches. Being mentioned in quality sources is one of the levers for influencing their generated content.

A search engine indexes the web and returns a list of relevant links. An LLM generates a synthetic response in natural language. The two are converging: Google integrates LLMs into its results, and LLMs like Perplexity incorporate real-time web searches.

No. Each developer builds its own corpus with different sources, cutoff dates, and filtering methods. This is why a brand may be well represented in ChatGPT but absent from Claude, or vice versa. Multi-LLM monitoring is essential.

Related articles

SEO Comparison Pages: X vs Y
16 March 2026
-
SEO

Comparison Page SEO: How to Rank for "X vs Y" Queries

"X vs Y" queries capture users at the decision stage. Here's how to structure and optimize your comparison pages to dominate them in SEO.

Gemini, ChatGPT, Perplexity, Claude, Copilot differences?
10 March 2026
-
GEO / AEO

Gemini vs ChatGPT vs Perplexity vs Claude vs Copilot: What Are the Differences for Your Visibility in 2026?

Not all AIs cite content the same way. Here's what that changes for your visibility in 2026.

RAG (Retrieval-Augmented Generation)
09 March 2026
-
GEO / AEO

What Is RAG (Retrieval-Augmented Generation)? And Why It Impacts Your Visibility

RAG is why ChatGPT cites some websites and not others. Here's how this mechanism works, and what it means for you.

Related terms

GEO / AEO

GEO (Generative Engine Optimization)

GEO refers to all practices aimed at optimizing a brand's visibility in responses generated by LLMs such as ChatGPT, Claude, or Perplexity. The goal is not only to drive direct traffic, but to be cited as a reliable reference in AI-generated answers.

GEO / AEO

RAG (Retrieval-Augmented Generation)

RAG is a technique that extends an LLM by allowing it to retrieve information in real time from external sources before generating its response. Perplexity and Bing Copilot operate on this principle. It's what enables an LLM to cite up-to-date sources.

GEO / AEO

Query fan-out

Query fan-out is the mechanism by which an LLM breaks down a complex question into multiple complementary sub-queries to interrogate different sources before synthesizing a final response. It is a core mechanism of RAG systems like Perplexity.