CivBench: A Long-Horizon Benchmark for Tool-Mediated Agents in Civilization VI
CivBench is presented, an open-source benchmark for evaluating language model agents in long-horizon, tool-mediated environments through the Model Context Protocol (MCP), and two interface-level metrics are introduced that the environment makes measurable: Proactive Monitoring Rate (PMR) and RAG@10, capturing whether c...
Austin Andrews, L. Wilkinson, Jamie Heagerty et al.
· 0 citations