← all datasets

Humanity's Last Exam

Emerging

3papers using it

2025first seen

'Humanity's Last Exam' is a benchmark dataset used to evaluate the capabilities of evolving agents in scientific inquiry and experimentation.

🔎 Find this dataset

Papers using Humanity's Last Exam (3)

EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale2026

Cognitive Loop via In-Situ Optimization: Self-Adaptive Reasoning for Science2025

SciMaster: Towards General-Purpose Scientific AI Agents, Part I. X-Master as Foundation: Can We Lead on Humanity's Last Exam?2025

Humanity's Last Exam — datasets — ai-for-science