OS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Zhenyu, Xie, Jingjing, Li, Zehao, Yang, Bowen, Sun, Qiushi, Liu, Zhaoyang, Liu, Zhoumianze, Qiao, Yu, Yue, Xiangyu, Wang, Zun, Ding, Zichen
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914207130189824
author Wu, Zhenyu
Xie, Jingjing
Li, Zehao
Yang, Bowen
Sun, Qiushi
Liu, Zhaoyang
Liu, Zhoumianze
Qiao, Yu
Yue, Xiangyu
Wang, Zun
Ding, Zichen
author_facet Wu, Zhenyu
Xie, Jingjing
Li, Zehao
Yang, Bowen
Sun, Qiushi
Liu, Zhaoyang
Liu, Zhoumianze
Qiao, Yu
Yue, Xiangyu
Wang, Zun
Ding, Zichen
contents With VLM-powered computer-using agents (CUAs) becoming increasingly capable at graphical user interface (GUI) navigation and manipulation, reliable step-level decision-making has emerged as a key bottleneck for real-world deployment. In long-horizon workflows, errors accumulate quickly and irreversible actions can cause unintended consequences, motivating critic models that assess each action before execution. While critic models offer a promising solution, their effectiveness is hindered by the lack of diverse, high-quality GUI feedback data and public critic benchmarks for step-level evaluation in computer use. To bridge these gaps, we introduce OS-Oracle that makes three core contributions: (1) a scalable data pipeline for synthesizing cross-platform GUI critic data; (2) a two-stage training paradigm combining supervised fine-tuning (SFT) and consistency-preserving group relative policy optimization (CP-GRPO); (3) OS-Critic Bench, a holistic benchmark for evaluating critic model performance across Mobile, Web, and Desktop platforms. Leveraging this framework, we curate a high-quality dataset containing 310k critic samples. The resulting critic model, OS-Oracle-7B, achieves state-of-the-art performance among open-source VLMs on OS-Critic Bench, and surpasses proprietary models on the mobile domain. Furthermore, when serving as a pre-critic, OS-Oracle-7B improves the performance of native GUI agents such as UI-TARS-1.5-7B in OSWorld and AndroidWorld environments. The code is open-sourced at https://github.com/numbmelon/OS-Oracle.
format Preprint
id arxiv_https___arxiv_org_abs_2512_16295
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic Models
Wu, Zhenyu
Xie, Jingjing
Li, Zehao
Yang, Bowen
Sun, Qiushi
Liu, Zhaoyang
Liu, Zhoumianze
Qiao, Yu
Yue, Xiangyu
Wang, Zun
Ding, Zichen
Artificial Intelligence
With VLM-powered computer-using agents (CUAs) becoming increasingly capable at graphical user interface (GUI) navigation and manipulation, reliable step-level decision-making has emerged as a key bottleneck for real-world deployment. In long-horizon workflows, errors accumulate quickly and irreversible actions can cause unintended consequences, motivating critic models that assess each action before execution. While critic models offer a promising solution, their effectiveness is hindered by the lack of diverse, high-quality GUI feedback data and public critic benchmarks for step-level evaluation in computer use. To bridge these gaps, we introduce OS-Oracle that makes three core contributions: (1) a scalable data pipeline for synthesizing cross-platform GUI critic data; (2) a two-stage training paradigm combining supervised fine-tuning (SFT) and consistency-preserving group relative policy optimization (CP-GRPO); (3) OS-Critic Bench, a holistic benchmark for evaluating critic model performance across Mobile, Web, and Desktop platforms. Leveraging this framework, we curate a high-quality dataset containing 310k critic samples. The resulting critic model, OS-Oracle-7B, achieves state-of-the-art performance among open-source VLMs on OS-Critic Bench, and surpasses proprietary models on the mobile domain. Furthermore, when serving as a pre-critic, OS-Oracle-7B improves the performance of native GUI agents such as UI-TARS-1.5-7B in OSWorld and AndroidWorld environments. The code is open-sourced at https://github.com/numbmelon/OS-Oracle.
title OS-Oracle: A Comprehensive Framework for Cross-Platform GUI Critic Models
topic Artificial Intelligence
url https://arxiv.org/abs/2512.16295