BEPA-7B-S2

From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy Assimilation

🌐 Project Page | 📑 arXiv Paper | 💻 GitHub

🏆 #1 Open-Source End-to-End Model on OSWorld (15 steps): Achieves 32.13% success rate
📊 Extreme Data Efficiency: Matches GUI-OWL-7B performance using only 128 training tasks

Model Description

BEPA-7B-S2 is a GUI agent model fine-tuned from UI-TARS-1.5-7B using the BEPA (Bi-Level Expert-to-Policy Assimilation) framework. This model achieves state-of-the-art performance among open-source end-to-end models on the OSWorld benchmark.

Key Results

Method	D_{expert_only}	D_train	D_{held_out}	Overall (%)
UITARS1.5-7B	18.52	55.12	5.74	22.87
GRPO	11.11	58.02	5.32	23.60
BEPA (ours)	35.19	73.23	10.30	32.13

BEPA improves UI-TARS-1.5-7B from 22.87% to 32.13% on OSWorld-Verified (+9.26 points, +40.5% relative improvement).

BEPA Framework

BEPA Overview

BEPA addresses two key challenges when using expert trajectories for training end-to-end GUI policies:

Structural Mismatch: Framework traces interleave multiple roles (planning, execution, grounding) that end-to-end policies cannot directly imitate.
Distribution Gap: Even after format conversion, trajectories remain far from the base-policy manifold.

LEVEL-1: Self-Rolled Execution

Transforms alien expert traces into policy-compatible trajectories by abstracting expert trajectories into compact natural-language plans, then letting the base policy act in the environment with plan conditioning.

LEVEL-2: Self-Aligned Assimilation

Dynamically maintains a per-task cache, injecting guided trajectories into GRPO updates only upon total on-policy failure. The cache is continuously refreshed with the policy's own successful executions.

Citation

@misc{wang2026offpolicyonpolicyenhancinggui,
      title={From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy Assimilation},
      author={Zezhou Wang and Ziyun Zhang and Xiaoyi Zhang and Zhuzhong Qian and Yan Lu},
      year={2026},
      eprint={2601.05787},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2601.05787},
}