What is LLM (Large Language Model)?
An LLM (Large Language Model) is an AI model trained on enormous amounts of text from the internet that understands and generates natural language by predicting the most probable next word. GPT, Gemini, Claude, Llama and DeepSeek are well-known examples.
“Large” refers to the very high number of parameters (the model's internal tunable values) and the sheer volume of training data. As models have scaled up, they have become able to handle a wide range of tasks such as translation, summarization, reasoning and coding.
AI chatbot services are built on top of these LLMs, combining them with features like web search, file reading and voice. The choice of which LLM a service uses shapes much of its performance and character.