# What is Foundation Model? Definition & Technical Mechanics

> **Source:** [SPIDITS AI AI Glossary](https://spidits.com/ai-glossary)  
> **Category:** Foundational AI  
> **Canonical Citation:** [https://spidits.com/ai-glossary/foundation-model](https://spidits.com/ai-glossary/foundation-model)

## Definition
A Foundation Model is a large-scale AI model trained on massive, broad datasets (typically through self-supervised learning) that serves as the baseline starting point for multiple downstream tasks. Examples include GPT-4, LLaMA, and stable diffusion models.

## Architectural Mechanics & Deep Dive
A foundation model is a large-scale AI model trained on vast quantities of unlabeled data (usually via self-supervised learning) that can be adapted to a wide range of downstream tasks. Examples include GPT-4, Claude, and Llama. These models serve as a general cognitive starting point, allowing developers to fine-tune or prompt them for diverse applications without training from scratch.

## Industry Applications & Real-World Use Cases
Base model development, specialized API features, and downstream fine-tuning

## Frequently Asked Questions

### Why are they called foundation models?
Because they represent a solid linguistic or visual foundation that developers can specialize for thousands of separate specific apps without training from scratch.

### Who coined the term Foundation Model?
The Stanford Institute for Human-Centered Artificial Intelligence (HAI) in 2021.

## Related Terminology
- [llm](https://spidits.com/ai-glossary/llm)
- [llama](https://spidits.com/ai-glossary/llama)
- [generative-ai](https://spidits.com/ai-glossary/generative-ai)

---
*From the SPIDITS AI Knowledge Base (336+ terms indexed). Explore full technical definitions: [https://spidits.com/ai-glossary](https://spidits.com/ai-glossary)*