Apple has taken a significant step by opening its foundational AI model to third-party developers for the first time. This development, announced during the recent Worldwide Developers Conference, marks a shift in Apple’s traditionally closed approach to AI tools, enabling direct access to the on-device large language model that powers Apple Intelligence. The new model, featuring three billion parameters, operates solely on-device.
This aligns with Apple’s commitment to user privacy, though it does come with technical limitations compared to cloud-based solutions offered by competitors. Craig Federighi, Apple’s software chief, highlighted this development, stating that any app can now tap directly into the large language model core. With the introduction of the Foundation Models framework, developers can easily integrate Apple Intelligence capabilities using just three lines of Swift code.
This framework emphasizes privacy-focused AI inference at no cost and includes built-in guided generation and tool-calling functionalities. Notably, Automattic, the company behind the Day One journaling app, has already begun utilizing this framework to enhance user experience. Moreover, Apple has integrated large language models into Xcode 26, allowing developers to use ChatGPT seamlessly without needing to create an account.
This update also includes advanced coding tools that suggest actions such as generating previews or resolving code issues, enhancing the development process. Alongside these features, Apple has extended its Visual Intelligence capabilities to third-party developers. Etsy is exploring these visual tools for product discovery, showcasing the potential for enhanced customer engagement through camera-based searches.
Despite these advancements, market analysts have responded with skepticism, noting that the features presented seemed incremental, particularly given the competitive AI landscape. Apple’s focus appears to be on building a robust AI infrastructure grounded in privacy, rather than making bold public claims. As testing for Apple’s AI developer tools begins, the company seems committed to a measured approach that prioritizes privacy and sustainable innovation over flashy advancements, indicating a careful strategy as it navigates the evolving AI landscape.