An LLM (Large Language Model) is a neural network trained on enormous amounts of text (and often other modalities) to predict the next token. That ability lets it summarise, answer, translate, code, and reason at human-readable quality. GIGI uses LLMs internally; you generally do not pick the model — you pick the behaviour you want, and we route to the right one.
AI agents
LLM
A Large Language Model — the model powering chat, voice, and agent reasoning.
See also
- AI agent
Software that takes goals, plans steps, uses tools, and acts on your behalf.
- Context window
How much text/history the agent can see at once.
- Inference
Actually running the AI model to produce an answer.
- Fine-tuning
Extra training so a model speaks in your brand's voice or knows your specific data.