OpenCodePapers

computer-use-agents-on-windowsagentarena-15-steps

AgentsComputer Use Agents
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeSuccess RateModelNameReleaseDate
Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents✓ Link35.9RoTS-32B2026-05-28
Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis✓ Link30.17JEDI-7B w/ GPT-4o2025-05-19
Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents✓ Link29.8Agent S2 w/ Claude-3.7-Sonnet2025-04-01
Watch and Learn: Learning to Use Computers from Online Videos24.0UI-TARS-1.5-7B (W&L SFT)2025-10-06
UltraCUA: A Foundation Model for Computer Use Agents with Hybrid Action21.7UltraCUA-7B2025-10-20
CoAct-1: Computer-using Multi-Agent System with Coding Actions21.4CoAct-12025-08-05
ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data✓ Link21.4ScaleCUA-32B2025-09-18
Windows Agent Arena: Evaluating Multi-Modal OS Agents at Scale✓ Link19.5Navi w/ GPT-4V-1106 (UIA + OmniParser)2024-09-12
Agent S: An Open Agentic Framework that Uses Computers Like a Human✓ Link18.2Agent S w/ GPT-4o2024-10-10
Check out how to contribute new results. Then edit this file on GitLab or GitHub.