Large language model (LLM)
A large language model is an AI system trained on very large amounts of text to predict the next piece of text. That simple skill lets it draft, summarize, translate, classify and work through problems. Claude, GPT, Gemini and Grok are examples of LLM families. ChatGPT and similar apps are products built on top of them.
What it is
An LLM is a neural network trained on large collections of text, and often images and audio too. It works in tokens, small chunks of text. For text generation, it predicts successive tokens using the material supplied in its context window. IBM groups LLMs under generative AI: models that create new content in response to a prompt.
How it works
Two phases matter for business users:
| Phase | What happens | Who does it | How often |
| — | — | — | — |
| Training | Pretraining learns patterns; later training can adapt behavior or skills. A published knowledge cutoff describes training-data recency, not guaranteed knowledge of every earlier fact. | The model developer or a team adapting it | Multiple stages; releases and updates vary |
| Use (inference) | You send a prompt and the model predicts a response. Apps add memory, search and tools around it. | You, every request | Every time |
Vendors publish model lists with context sizes, cutoffs and prices. Anthropic and OpenAI both document theirs.
Benefits
-
General-purpose: one model handles writing, analysis, coding and more.
-
Works in plain language, so non-technical staff can use it.
-
Can be connected to your data and tools to do real work.
Limits and risks
-
Can state false things confidently (see hallucination).
-
A cutoff is not live access or a completeness guarantee. Supply current evidence for changing facts; search and documents can also contain errors.
-
Output varies between runs and between models.
Agent compatibility
Not tool-specific. See Agent-ready for connector support by tool.
Levels come from each tool's recorded, sourced compatibility. How the levels work.
Related entries
Sources
- What is generative AI? · IBMreference accessed Oct 17, 2026
- Models overview · Anthropicofficial accessed Oct 17, 2026
- Models · OpenAIofficial accessed Oct 17, 2026
Practitioner views are restated in our words and labeled as views, not measured results. Product facts come from official sources dated above.
Suggest an edit
IonBin's wiki isn't user-edited. Corrections go to our Directory Curator, who checks them against primary sources before anything changes. Spot something wrong or out of date on “Large language model (LLM)”? Email us with a link to a primary source.
Suggest an edit by email or write to contact@ionbin.com