# Large language model

A learned model that produces or scores token sequences from context; in an agent, it is the reasoning component rather than the complete system.

technical-term · Models & inference · foundational · Reviewed 2026-08-09

## Definition

A large language model estimates language-conditioned outputs from a context of tokens. It may generate text, structured data, code, or tool-call arguments, but it does not by itself provide identity, permission, durable memory, execution, verification, evidence, or accountability.

The model class is established and rapidly evolving; capabilities and limitations differ materially by model, configuration, and task.

## Why it matters

Model behavior is probabilistic and context-dependent.

Enterprise claims must be measured on the complete deployed system and workload, not inferred from a public leaderboard.

## System anatomy

- **Parameters:** Learned weights encoding statistical patterns.
- **Context:** The finite token sequence supplied at inference time.
- **Decoder:** The process that selects output tokens under configured sampling.



## Important distinctions

- **AI agent:** A model generates candidate decisions; an agent loops, uses tools, maintains state, and operates inside a governed system.
- **Knowledge base:** Model weights are not a current, attributable store of enterprise truth.



## Implementation signals

- Select models through workload-specific evaluations
- Record model identity, version, configuration, and routing decisions

## Failure modes

- Treating fluency as factual authority
- Assuming one benchmark ranking predicts production fitness



## Related knowledge

- [AI agent](https://aisdlc.ai/agentic-engineering/ai-agent) — A goal-directed software system in which a model dynamically chooses steps and tools, observes results, updates working state, and continues until an exit condition or human handoff.
- [Context window](https://aisdlc.ai/agentic-engineering/context-window) — The finite sequence of tokens a model can directly condition on during a single inference operation.
- [Model routing](https://aisdlc.ai/agentic-engineering/model-routing) — The policy-based selection of a model and configuration for a request, task stage, risk level, or fallback condition.
- [Mixture of experts](https://aisdlc.ai/agentic-engineering/mixture-of-experts) — A neural architecture that routes each input token or example through a selected subset of learned expert subnetworks.

## Sources and further study

- [Artificial Analysis — Artificial Analysis Intelligence Benchmarking Methodology](https://artificialanalysis.ai/methodology/intelligence-benchmarking) — A transparent methodology for comparing model quality, price, latency, throughput, and cost per task. Its composite indices are useful evidence, not a substitute for use-case-specific evaluation. **Use:** Independent benchmark. **Limitation:** Composite benchmark results change over time and do not establish enterprise fit, safety, controllability, or acceptance for a specific workload. No leaderboard values are reproduced here.
- [NIST — Towards best practices for automated benchmark evaluations](https://www.nist.gov/news-events/news/2026/01/towards-best-practices-automated-benchmark-evaluations) — Guidance on the integrity and limitations of automated benchmark evaluation practices. **Use:** Official guidance. **Limitation:** Guidance describes recommended practice; citation does not prove that a control is implemented or effective in a particular environment.
- [OpenAI — A practical guide to building agents](https://openai.com/business/guides-and-resources/a-practical-guide-to-building-ai-agents/) — Defines a practical agent around a model, tools, instructions, and a run loop that continues until an exit condition is reached. **Use:** Primary definition. **Limitation:** The guide is first-party product guidance and a simplified starting architecture, not a complete regulated-enterprise control model.

---

This library synthesizes cited research, standards, official documentation, and clearly attributed practitioner perspectives. Maturity describes the state of a concept—not vendor endorsement, production readiness, or permission to deploy. Benchmarks and demonstrations do not replace use-case evaluation, governed controls, independent verification, or named human release authority.