Search papers, labs, and topics across Lattice.
This paper introduces GUI-RobustEval, a benchmark with 1,216 executable test cases, to systematically evaluate GUI agents' error recovery capabilities. To improve robustness, they propose Robustness-driven Trajectory Synthesis (RoTS), a framework for generating 800k high-quality training examples that proactively discover diverse error modes and synthesize recovery steps. Fine-tuning 7B and 32B models on the RoTS dataset leads to significant performance gains on both GUI-RobustEval and OSWorld, demonstrating the importance of error recovery for overall performance.
GUI agents can now recover from errors far more effectively, thanks to a new benchmark and data synthesis method that specifically targets robustness.
While GUI agents have advanced rapidly, they often lack the robustness to recover from their own errors, hindering real-world deployment. To bridge this gap at both the evaluation and data levels, we introduce GUI-RobustEval and propose Robustness-driven Trajectory Synthesis. GUI-RobustEval contains $1,216$ executable test cases that systematically measure error recovery capabilities across a broad and realistic spectrum of error modes. At the data level, RoTS is a scalable synthesis framework that creates $800k$ high-quality data via a tree-based pipeline that proactively discovers diverse error modes and synthesizes corresponding recovery steps. Our two models, RoTS-7B and RoTS-32B, fine-tuned on our dataset, both demonstrate significant gains on GUI-RobustEval and traditional GUI benchmarks. Notably, RoTS-32B achieves state-of-the-art performance on OSWorld, with a $47.4\%$ success rate and a $33.8\%$ All-Pass@4 score, suggesting that improved long-horizon error recovery ability contributes to both robustness and overall performance. Our code is available at https://github.com/AlibabaResearch/RoTS.
