Back to Glossary
BasicsAI Glossary

Model

Quick Answer

A model is the trained output of a machine learning process — a collection of learned parameters that, combined with an algorithm, can turn new inputs into predictions or generated content without being explicitly programmed for each case.

In Depth

What Model really means

When people talk about 'the AI', they usually mean the model. A model encodes what the system has learned from its training data. Once trained, a model can be deployed as an API, embedded into software, or downloaded and run locally, depending on the use case and sensitivity of the data.

Models vary enormously in size and cost. A compact classification model might be a few megabytes; a frontier large language model may be hundreds of gigabytes and require specialist hardware to serve at scale.

Why It Matters

Business relevance for UK organisations

Choosing the right model size is a classic cost/quality tradeoff. Many UK SMEs over-invest in the largest available model when a smaller, fine-tuned one would deliver 95% of the value at 5% of the cost.

Real-world example

How this shows up in practice

A Sheffield SaaS company replaced a general-purpose LLM with a fine-tuned smaller model for support ticket classification, reducing inference cost by 83%.