Reproduce “LOLPlay: Benchmarking Real-Time Visual Agents in League of Legends” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zhaojian Yu
LOLPlay evaluates visual agents controlling League of Legends champions through screenshots and mouse-keyboard actions during uninterrupted play. Its 36-task suite checks champion-specific state and event requirements, measuring complete-task and partial subgoal success. Across 324 evaluations, GPT-6 Astra reaches 25.0% task success at low and medium effort, and 40.3% subgoal success at medium effort. Results describe configurations, not League-wide performance or latency causation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.