Meet the Major Foundation Models

A handful of companies build the foundation models that power almost everything in this stack. You don't need a technical breakdown of each one, but knowing the players helps you understand what's actually running under the tools you use.

OpenAI's GPT models power ChatGPT and are licensed into a huge number of third-party apps and plugins, which is part of why GPT-based tools are everywhere.

Anthropic's Claude models power Claude.ai and a growing list of coding and business tools, generally known for careful, longer-form reasoning.

Google's Gemini models power Google's own assistant and are tightly woven into Google's other products: Docs, Search, Android.

Meta's Llama models are open-weight, meaning other companies and developers can download and build on them directly, which is why you'll find Llama quietly powering tools that never mention its name.

If you want to see foundation models side by side without committing to one, sites like Abacus.AI let you sample several in one place and compare raw output on the same prompt. That's about as close to "Level 1" as most people ever need to get, and it's genuinely useful for seeing that these models aren't interchangeable: they have real differences in tone, accuracy, and strengths even before anyone builds a product on top of them.