OSWorld

The 'OSWorld' dataset is a benchmark used to evaluate GUI automation tasks and tool-calling tasks for open-source models.

Papers using OSWorld (23)

Learn from Weaknesses: Automated Domain Specialization for Small Computer-Use Agents2026Mobile-agent-v3.5: Multi-platform Fundamental GUI Agents2026OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent2026Windows Agent Arena: Evaluating Multi-modal OS Agents At Scale2024 · 2 citesInfantagent-next: A Multimodal Generalist Agent For Automated Computer Interaction2025Learning from Failure: Inference-Time Self-Improvement for Computer-Use Agents2026Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents2026IntentScore: Intent-Conditioned Action Evaluation for Computer-Use Agents2026Agent Alpha: Tree Search Unifying Generation, Exploration and Evaluation for Computer-Use Agents2026Agent Skills for Large Language Models: Architecture, Acquisition, Security, and the Path Forward2026InfiniteWeb: Scalable Web Environment Synthesis for GUI Agent Training2026BEAP-Agent: Backtrackable Execution and Adaptive Planning for GUI Agents2026Surfer 2: The Next Generation Of Cross-platform Computer Use Agents2025Scaling Agents For Computer Use2025Agentic Lybic: Multi-Agent Execution System with Tiered Reasoning and Orchestration2025Mano Technical Report2025Instruction Agent: Enhancing Agent With Expert Demonstration2025UI-TARS-2 Technical Report: Advancing GUI Agent With Multi-turn Reinforcement Learning2025CoAct-1: Computer-using Multi-Agent System with Coding Actions2025OSWorld-Human: Benchmarking the Efficiency of Computer-Use Agents2026Ui-evol: Automatic Knowledge Evolving For Computer Use Agents2025Osworld: Benchmarking Multimodal Agents For Open-ended Tasks In Real Computer Environments2024 · 9 citesAgent S: An Open Agentic Framework that Uses Computers Like a Human2024 · 2 cites