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

• URL: http://arxiv.org/abs/2510.23691v1
• Date: Mon, 27 Oct 2025 17:43:51 GMT
• Title: Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents
• Authors: Zihao Wang, Xujing Li, Yining Ye, Junjie Fang, Haoming Wang, Longxiang Liu, Shihao Liang, Junting Lu, Zhiyong Wu, Jiazhan Feng, Wanjun Zhong, Zili Li, Yu Wang, Yu Miao, Bo Zhou, Yuanfan Li, Hao Wang, Zhongkai Zhao, Faming Wu, Zhengxuan Jiang, Weihao Tan, Heyuan Yao, Shi Yan, Xiangyang Li, Yitao Liang, Yujia Qin, Guang Shi,
• Abstract summary: We present Game-TARS, a generalist game agent trained with a unified, scalable action space anchored to human-aligned keyboard-mouse inputs.<n>Game-TARS is pre-trained on over 500B tokens with diverse trajectories and multimodal data.<n> Experiments show that Game-TARS achieves about 2 times the success rate over the previous sota model on open-world Minecraft tasks.
• Score: 56.25101378553328
• License: http://creativecommons.org/publicdomain/zero/1.0/
• Abstract: 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.

Related papers

• NitroGen: An Open Foundation Model for Generalist Gaming Agents [101.41866522979548]
  NitroGen is a vision-action foundation model for generalist gaming agents.<n>It is trained on 40,000 hours of gameplay videos across more than 1,000 games.
  arXiv  Detail & Related papers  (2026-01-04T16:24:50Z)
• Learning Controllable and Diverse Player Behaviors in Multi-Agent Environments [0.0]
  This paper introduces a reinforcement learning framework that enables controllable and diverse player behaviors without relying on human gameplay data.<n>We define player behavior in an N-dimensional continuous space and uniformly sample target behavior vectors from a region that encompasses a subset representing real human styles.<n>A single PPO-based multi-agent policy can reproduce new or unseen play styles without retraining.
  arXiv  Detail & Related papers  (2025-12-11T17:26:24Z)
• Learning Representations in Video Game Agents with Supervised Contrastive Imitation Learning [0.6299766708197881]
  This paper introduces a novel application of Supervised Contrastive Learning (SupCon) to Imitation Learning (IL)<n>The goal is to obtain latent representations of the observations that capture better the action-relevant factors.<n> Experiments on the 3D games Astro Bot and Returnal, and multiple 2D Atari games show improved representation quality, faster learning convergence, and better generalization.
  arXiv  Detail & Related papers  (2025-09-15T13:00:29Z)
• Scaling Offline Model-Based RL via Jointly-Optimized World-Action Model Pretraining [49.730897226510095]
  We introduce JOWA: Jointly-Reinforced World-Action model, an offline model-based RL agent pretrained on Atari games with 6 billion tokens data.<n>Our largest agent, with 150 million parameters, 78.9% human-level performance on pretrained games using only 10% subsampled offline data, outperforming existing state-of-the-art large-scale offline RL baselines by 31.6% on averange.
  arXiv  Detail & Related papers  (2024-10-01T10:25:03Z)
• Sports-Traj: A Unified Trajectory Generation Model for Multi-Agent Movement in Sports [53.637837706712794]
  We propose a Unified Trajectory Generation model, UniTraj, that processes arbitrary trajectories as masked inputs.<n>Specifically, we introduce a Ghost Spatial Masking (GSM) module, embedded within a Transformer encoder, for spatial feature extraction.<n>We benchmark three practical sports datasets, Basketball-U, Football-U, and Soccer-U, for evaluation.
  arXiv  Detail & Related papers  (2024-05-27T22:15:23Z)
• Scaling Laws for Imitation Learning in Single-Agent Games [28.257046559127875]
  We investigate whether carefully scaling up model and data size can bring similar improvements in the imitation learning setting for single-agent games.<n>We first demonstrate our findings on a variety of Atari games, and thereafter focus on the extremely challenging game of NetHack.<n>We find that IL loss and mean return scale smoothly with the compute budget and are strongly correlated, resulting in power laws for training compute-optimal IL agents.
  arXiv  Detail & Related papers  (2023-07-18T16:43:03Z)
• Multi-Game Decision Transformers [49.257185338595434]
  We show that a single transformer-based model can play a suite of up to 46 Atari games simultaneously at close-to-human performance. We compare several approaches in this multi-game setting, such as online and offline RL methods and behavioral cloning. We find that our Multi-Game Decision Transformer models offer the best scalability and performance.
  arXiv  Detail & Related papers  (2022-05-30T16:55:38Z)
• Deep Policy Networks for NPC Behaviors that Adapt to Changing Design Parameters in Roguelike Games [137.86426963572214]
  Turn-based strategy games like Roguelikes, for example, present unique challenges to Deep Reinforcement Learning (DRL) We propose two network architectures to better handle complex categorical state spaces and to mitigate the need for retraining forced by design decisions.
  arXiv  Detail & Related papers  (2020-12-07T08:47:25Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.