Game-TARS：面向可扩展通用型多模态游戏智能体的预训练基础模型

摘要

我们提出Game-TARS——一种基于统一可扩展动作空间的通用游戏智能体，其动作空间以符合人类习惯的键盘鼠标原生输入为锚点。与基于API或图形界面的方法不同，该范式支持跨操作系统、网页和模拟游戏等异构领域的大规模持续预训练。Game-TARS通过5000亿标记的多模态数据及多样化轨迹进行预训练，核心技术包括降低因果混淆的衰减持续损失函数，以及平衡推理深度与计算成本的稀疏思维策略。实验表明：在开放世界《我的世界》任务中，Game-TARS的成功率达到此前最优模型的约两倍；在未见过的网页3D游戏中接近人类新手的普适性水平；在FPS游戏基准测试中超越GPT-5、Gemini-2.5-Pro和Claude-4-Sonnet。训练阶段与测试阶段的扩展性实验证实，统一动作空间在跨游戏多模态数据扩展时能持续提升性能。我们的结果表明：简洁可扩展的动作表征与大规模预训练相结合，为构建具有广泛计算机使用能力的通用智能体提供了可行路径。

English

We present Game-TARS, a generalist game agent trained with a unified, scalable action space anchored to human-aligned native keyboard-mouse inputs. Unlike API- or GUI-based approaches, this paradigm enables large-scale continual pre-training across heterogeneous domains, including OS, web, and simulation games. Game-TARS is pre-trained on over 500B tokens with diverse trajectories and multimodal data. Key techniques include a decaying continual loss to reduce causal confusion and an efficient Sparse-Thinking strategy that balances reasoning depth and inference cost. Experiments show that Game-TARS achieves about 2 times the success rate over the previous sota model on open-world Minecraft tasks, is close to the generality of fresh humans in unseen web 3d games, and outperforms GPT-5, Gemini-2.5-Pro, and Claude-4-Sonnet in FPS benchmarks. Scaling results on training-time and test-time confirm that the unified action space sustains improvements when scaled to cross-game and multimodal data. Our results demonstrate that simple, scalable action representations combined with large-scale pre-training provide a promising path toward generalist agents with broad computer-use abilities.

Game-TARS：面向可扩展通用型多模态游戏智能体的预训练基础模型

Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents

摘要

Support