LLM Reference
Agent

ADRS (UCB-ADRS)

About

Systems problems benchmark for evaluating agentic AI coherence, persistence, and reliability in complex multi-step tasks.