The Foundation Models Framework: Diving Into Apple Intelligence

The Foundation Models Framework: Diving Into Apple Intelligence

From A-Positive by Studio Rubio

January 18, 2026 · 22 min · Episode 11

About this episode

The episode discusses Apple's Foundation Models Framework, which introduces a large language model for on-device AI processing.

Apple’s new Foundation Models Framework introduces an on-device large language model (LLM) with 3 billion parameters, enabling AI features to run locally on Apple devices. This “device-scale model” excels at tasks like content generation, text summarization, and natural language understanding, all while prioritizing user privacy by processing data locally and offline. The framework also introduces Guided Generation, allowing developers to define the structure of AI-generated responses, ensuring reliability and robustness in app features. Links: Design and Development Videos - Apple Developer Apple Just Handed Every Developer a 3-Billion Parameter AI Model, No Cloud Required | by Cogni Down Under | Medium Apple’s Foundation Models framework unlocks new intelligent app experiences - Apple Guided Generation with FoundationModels: How to Get Swift Structs from LLMs | by Luiz Fernando Salvaterra | Medium Deep dive into the Foundation Models framework - WWDC25 - Videos - Apple Developer Introduction to Apple's FoundationModels: Limitations, Capabilities, Tools Apple supercharges its tools and technologies for developers - Apple Joint statement from Google and Apple ‎Local AI Chat…

Topics covered

  • Apple
  • Foundation Models Framework
  • AI
  • large language model
  • user privacy

Keywords

  • content generation
  • text summarization
  • natural language understanding
  • Guided Generation

Mentioned in this episode

Products: The Foundation Models Framework: Diving Into Apple Intelligence, Apple Foundation Models Framework

More episodes of A-Positive

Explore listener stats, chart rankings, contacts and more on the A-Positive podcast page.