Comparing Fable and 10 LLMs on refactoring a LangGraph god node

An experiment compared 11 LLMs (5 American, 6 Chinese) on refactoring a LangGraph god node. Models generated proposals and evaluated each other, with full results and ranking scripts published.