Rohan Paul (@rohanpaul_ai) · 10/11/2026, 21:02:00
Tsinghua AAArena Benchmark Shows AI Agents Can Top Game Leaderboards via Replay Analysis but Stall on Complex Rules
Tsinghua University introduced the AAArena benchmark to evaluate heuristic learning in AI coding agents within long-running game competitions. The study reveals that detailed match replay feedback significantly outperforms win/loss-only signals, propelling a Pacman bot from rank 11 to rank 1; however, tripling the match budget failed to help stuck agents overcome limitations in games with complex rules.
SOURCE COVERAGEFull Original Report
Rohan Paul (@rohanpaul_ai) · 2026-10-11 13
New Tsinghua paper finds that AI agents improving game bots from match replays can top human leaderboards, but mostly stall on games with complex rules.
Getting AI to learn a winning game strategy from a limited number of matches is still hard, especially against changing rivals.
They built AAArena from 12 games in Tsinghua's yearly bot-building contest, with 1,920 archived human programs as rivals. A coding agent, with its model weights unchanged, reads the rules, picks opponents, studies replays, and rewrites its bot within a match budget.
Detailed replays beat win/loss-only feedback in all 3 games tested. With replays, a Pacman bot reached rank 1, versus rank 11 without them.
Tripling the match budget did not push any of 4 stuck bots to rank 1.
– arxiv. org/abs/2610.12341
Title: "Can AI Agents Learn Their Way to the Top? Evaluating Heuristic Learning in a Long-Running Game Agent Competition"
