cs.AINov 18, 2026

Think Offline, Act in Real Time: Self-Improving Visual Game Agents through Code

Authors: Zipeng Qiu, Shusheng Xu, Jiale Li, Yukun Zhou, Chenyue Li, Wenshuang Li, Tongkai Yang, Binhang Yuan

Organizations: Ant Group · HKUST

Abstract

Pre-trained models can read screens and reason about games, yet on an unfamiliar game, an agent still has to learn to play. We ask how an agent can self-improve on such a game from a few of its own attempts, in real-time visual games where the agent sees only the screen and the game keeps running while the agent thinks. We study two kinds of policy that can carry this improvement. In model as policy, a model selects actions during play; we adapt real-time chunking so that committed actions keep executing while the model reasons, and train a 2B controller for this interface. In code as policy, a model writes, tests, and revises a program between attempts, and the program acts during play. Across 170 GameWorld tasks, with at most eight one-minute executions per run, code as policy reaches 47.80% mean progress, against 30.90% for our trained controller, and leads in all five game categories. Improving the controller's weights on a single game, by contrast, yields only modest gains. These results favor letting a model think offline and storing what it learns in a program that acts in real time. Project repository: https://github.com/TabibitoQZP/SIGame

Figures & tables

Appendix figures & tables16 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Game2Player: Scaling Gameplay Experience for Visual Agents through Environment-Player Co-Development

    Nov 18, 2026Xirui Jiang, Yuxi Chen, Rui Yang +11Game-Playing AgentsAI Agent Benchmarks

  2. The World Won't Wait: Evaluating and Teaching Agents Under a Live Clock

    Oct 14, 2026Ningxin Pan, Hanyu Li, Yeqing Shen +3LLM Agent Self-ImprovementLLM Decision-Making