FM-Bench: A Benchmark for Long-Horizon Management with Competing Agents

FM-Bench, the Football Management Benchmark, evaluates whether LLM agents can sustain effective decision-making over long horizons with cumulative consequences. Agents manage a football club for 20 in-game years, utilizing 26 tools across 340 to 400 decision stops. Tasks include drafting squads, trading players, negotiating contracts, investing in youth facilities, and setting lineups. Performance is measured by a deterministic engine that calculates a final score, removing the need for human raters or LLM judges.