Humanity's Last Exam (HLE)

The 'Humanity's Last Exam (HLE)' dataset/benchmark is used to evaluate the reasoning capabilities of artificial intelligence by assessing its ability to formulate, adapt, and test thought patterns in dynamic conditions.

Papers using Humanity's Last Exam (HLE) (3)

EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale2026Cognitive Loop via In-Situ Optimization: Self-Adaptive Reasoning for Science2025SciMaster: Towards General-Purpose Scientific AI Agents, Part I. X-Master as Foundation: Can We Lead on Humanity's Last Exam?2025