Benchmarks are Not Enough: RAMP for Runtime Assessing of Agentic Models in Production Systems

Yipeng Ouyang, Xin Huang, Bingjie Liu +3

LLM agents are rapidly evolving from coding assistants into autonomous software engineering systems. However, existing evaluation methodologies remain largely centered on static, i…