Photo by MART PRODUCTION on Pexels

A recent study originating from Apple suggests that even the most sophisticated AI reasoning models, such as OpenAI’s o3, Anthropic’s Claude, and DeepSeek’s R1, fall short when confronted with highly complex problem-solving tasks. The findings, initially shared on Reddit (link: https://old.reddit.com/r/artificial/comments/1lvifsh/apples_new_study_shows_that_advanced_ai_reasoning/), shed light on the ongoing challenges in achieving true AI understanding and logical deduction.