| title | Recursive Environment Scaling for General Long-Horizon Agents |
|---|---|
| authors | Lingxiao Jing, Zhengbo Jiao, Yichen Xiang, Bofan Zhu, Lingbei Meng, Yan Ni, Haorui Chen, Ran Ran, Zhirong Shen, Yuheng Yang, Chenhang Cui, Yaosong Du, Rui Huang |
| date | 2026-10-12 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2610.16397 |
| upvotes | 6 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | We study environment scaling as a way to increase agent-benchmark difficulty while holding core task objectives fixed. Our framework, Recursive Environment Scaling (RES), transforms existing benchmark environments while preserving their objectives and completion requirements. A proposer changes information access, resource relations, and execution dependencies; execution-grounded checks validate the resulting candidate, and accepted changes are inherited across rounds. This produces related environments for paired evaluation and training-trajectory collection. Experiments on 294 initial tasks from three benchmarks show lower task success for all six evaluation models, with decreases of 57.14--77.14 percentage points on Claw-Eval-Live. With equal trajectory counts and identical optimization settings, fine-tuning on RES trajectories outperforms original-trajectory SFT in all nine model--benchmark comparisons by 7.00--14.61 percentage points. Ablations identify recursive inheritance and dependency composition as important contributors to increased difficulty. Together, these results support environment scaling for both agent evaluation and training. |
cite
@misc{jing2026recursive,
title={Recursive Environment Scaling for General Long-Horizon Agents},
author={Lingxiao Jing and Zhengbo Jiao and Yichen Xiang and Bofan Zhu and Lingbei Meng and Yan Ni and Haorui Chen and Ran Ran and Zhirong Shen and Yuheng Yang and Chenhang Cui and Yaosong Du and Rui Huang},
year={2026},
eprint={2610.16397},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2610.16397}
}