Reproduce “Gamer: A Generalist Vision-Language-Action Foundation Model for Long-Horizon Real-Time Game Agent” 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

Run Luo, Kuan Zhang, Qi Deng, Yukun Chen, Zhihao Yang, Yufei Zhang, Renrong Hu, Yi Zhang and 19 more

Gamer is an agent for long quests in Honkai: Star Rail, using keyboard-and-mouse control with task instructions and memory summaries. It combines continued pre-training, supervised fine-tuning, and online reinforcement learning across finite interaction segments, using game-verified outcomes. On two held-out quests in the same game, Gamer-RL raises milestone-based accuracy over Gamer-SFT; these are not full-quest completion rates.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.