A new study by expert network Pearl found that top frontier AI models like GPT-5.5 and Claude Opus 4.7 are only marginally better than older versions, with none exceeding a C+ grade. The models struggled with professional judgment, failing to prioritize or escalate when necessary, despite claims of superior intelligence. While some domains like business saw higher scores, law and health domains showed dangerously low expert alignment, suggesting human oversight remains critical. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT New research suggests current frontier AI models still lack the professional judgment and trustworthiness required for critical applications, indicating a need for continued human oversight.
RANK_REASON The cluster analyzes and critiques the performance of existing AI models based on a third-party study, rather than announcing a new release or significant industry event.
