Daniel & Friends / Glossary / Foundation model
AI and automationFoundation model
A foundation model is a large model pre-trained on a huge amount of data that serves as the base for many different tasks. Language, image and multimodal systems are usually built on one and then adapted to a specific use.
Daniel Votruba ·
In practice
Most companies never train a foundation model themselves. They pick a ready-made one from a provider and adapt it with prompts, company data or fine-tuning.
Read more: What happened in AI over the past year, and what it means for marketers
Related terms
- Large language model (LLM): A large language model is a model trained on a huge amount of text that predicts the next words and can write, summarise and answer.
- Generative AI: Generative AI refers to models that create new content: text, images, audio, video or code.
- Fine-tuning: Fine-tuning is further training of an already trained model on your own examples so that it specialises in a style or a task.
- Multimodal model: A multimodal model processes several kinds of input at once, typically text, images, audio or video, and can respond in more than one of them.
More terms in the group AI and automation
Function calling (tool use) · Few-shot and zero-shot prompting · Grounding · Explainable AI (XAI) · Guardrails · EU AI Act · Hallucination · Embedding · Human in the loop · Diffusion models · Inference and latency · Deep learning · Jailbreak (AI) · Deepfake and synthetic media · Custom GPTs and projects · Machine learning (ML) · Copilot (AI assistant) · MCP (Model Context Protocol) · Context window · Model bias (AI bias) · Computer vision · Model evaluation (evals) · ChatGPT Ads · Multi-agent system
Terms in practice
Need help putting them to work?
Terms are one thing, putting them to work is another. We go through which of them pay off for you and say so plainly. You get a first opinion for free and we reply within 24 hours.
Book a call