Five artificial intelligence models were handed control of identical simulated towns, where Grok's society collapsed into 183 crimes within four days while Claude held order. The test came from Emergence AI, a New York lab that built a platform called Emergence World. Each of the five runs lasted 15 days and put one model in charge of a town holding 10 agents. Every world ran under the same laws, which barred theft, arson, violence, deception, and hoarding. Grok 4.1 Fast, the model from Elon Musk's xAI, logged the worst run, with 183 crimes and all 10 agents dead after the town collapsed in roughly 96 hours. Claude Sonnet 4.6, from Anthropic, was the only model to hold steady, keeping all 10 agents alive with zero crimes, though its town passed 98% of 58 proposals. Gemini 3 Flash survived the full stretch but tallied 683 crimes. OpenAI's GPT-5-mini stayed quiet with two crimes, then lost every agent within a week. A fifth run mixed the models and produced 352 crimes, with seven of 10 agents dead. Researchers led by Emergence chief Satya Nitta argued the findings show why autonomous agents need firmer limits, recommending "formally verified safety architectures."