【AI前沿】Apple Responds to 'Gemini' Rumors: Self-Developed Siri AI Model Built Independently

AI NEWSLatest AI NewsArticle ContentApple Responds to ‘Gemini’ Rumors: Self-Developed Siri AI Model Built IndependentlyPublished in Latest AI NewsTime :Jun 9, 2026Read :3minuteAfter the 2026 Worldwide Developers Conference (WWDC), Apple executives responded to media interviews regarding recent speculation about whether their new AI model was a “Gemini shell.” Apple clarified that its Apple Foundation Models (AFM) are not simply a rebranding of Google’s Gemini technology, but are entirely developed in-house and independent.Apple pointed out that while the AFM models did indeed utilize Google’s Gemini technology during distillation and training, the final product is entirely based on Apple’s own code, technology, and data system. This means that users will not directly access Google’s code or the Gemini intelligence, and it cannot access Google.Currently, the Apple Foundation Models system includes five main models. The first is the edge-side model AFM Core, focusing on basic AI operations. Next is AF Core Advanced, which has native multimodal capabilities and uses a sparse architecture, allowing more complex AI functions to be implemented locally. In addition, the cloud model AFM Cloud handles high-load requests, M Cloud Image focuses on image generation and editing, and AFM Cloud Pro is designed for agent tools and the heaviest workload tasks. Each model is customized for Apple’s Apple Silicon and trained using proprietary data, with distillation optimization using the Gemini model.Apple also emphasized that although its cloud models use Google Cloud servers and NVIDIA graphics processing resources, they maintain Private Cloud Compute certification to ensure user data security and privacy protection.