MLE-Bench Lite

'MLE-Bench Lite' is a benchmark dataset used to evaluate the performance of machine learning systems in the context of long-horizon research engineering.

Papers using MLE-Bench Lite (3)

EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale2026Why Solve It Twice? Hierarchical Accumulation of Skills for Transfer-Efficient ML Engineering2026AceGRPO: Adaptive Curriculum Enhanced Group Relative Policy Optimization for Autonomous Machine Learning Engineering2026