Game-TARS：面向可擴展通用型多模態遊戲代理的預訓練基礎模型

摘要

我們提出Game-TARS——一種基於人類對齊的鍵鼠原生輸入錨定、採用統一可擴展動作空間訓練的通才型遊戲智能體。與基於API或GUI的方法不同，該範式支持跨異構領域的大規模持續預訓練，涵蓋操作系統、網頁及模擬遊戲等場景。Game-TARS通過5000億標記量的多模態數據與多樣化軌跡進行預訓練，其核心技術包括用於降低因果混淆的衰減持續損失函數，以及平衡推理深度與推斷成本的稀疏思維策略。實驗表明：Game-TARS在開放世界《我的世界》任務中達成約兩倍於前代最優模型的成功率，在未見網頁3D遊戲中接近人類新手的泛化能力，並在FPS基準測試中超越GPT-5、Gemini-2.5-Pro與Claude-4-Sonnet。訓練階段與測試階段的擴展實驗證實，統一動作空間在跨遊戲多模態數據擴展時能持續提升性能。我們的研究成果表明，簡潔可擴展的動作表徵與大規模預訓練相結合，為構建具備廣泛計算機使用能力的通才智能體開闢了可行路徑。

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：面向可擴展通用型多模態遊戲代理的預訓練基礎模型

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

摘要

Support