Apple announced new updates to its AI model at WWDC 2025, focusing on improved features across its intelligence suite. However, the tech giant’s recent models have faced scrutiny, with benchmark testing revealing a performance gap compared to rivals like OpenAI. While Apple emphasized advancements in on-device and server capabilities, enabling understanding of multiple languages and text generation, human testers rated their newest on-device model against Google and Alibaba’s offerings, highlighting only minor improvements. Apple Server, designed for data centers, lagged behind OpenAI’s GPT-4, while Meta’s Llama 4 performed better in image analysis. This divergence surprised Apple’s team, sparking skepticism from developers who feel the excitement around the event has diminished.