arXivAI Research
SciExam for ENSO: Evaluating AI Agents' Ability to Build Unsolved Climate Models
AI Agent 能建立氣候模型嗎？評估未知科學領域的「SciExam for ENSO」基準測試
The SciExam for ENSO benchmark evaluates AI agents tasked with building low-order climate models from scratch, revealing that half of the tested agents outperformed established human-published models.
2 min read