SpireBench shows agents lose the plot over long game runs
SpireBench gives 20 language models the full Slay the Spire game, exposing a long-horizon weakness that short benchmarks miss: frontier models match average players but remain far behind experts. Models often play the first act as well as experts, then fall behind because they misjudge risk across the run, saving health instead of spending it to grow stronger. Their written memories across runs are often wrong and overly cautious, and do not improve performance.