Apple Intelligence Foundation Language Models
This work addresses the need for efficient and private AI models for Apple users, but it is incremental as it applies existing methods to new data and platforms.
Apple developed foundation language models, including a ~3 billion parameter on-device model and a server-based model for Private Cloud Compute, to power Apple Intelligence features, achieving efficient and accurate performance across tasks with a focus on responsible AI.
We present foundation language models developed to power Apple Intelligence features, including a ~3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks efficiently, accurately, and responsibly. This report describes the model architecture, the data used to train the model, the training process, how the models are optimized for inference, and the evaluation results. We highlight our focus on Responsible AI and how the principles are applied throughout the model development.