OpenCodePapers
computer-use-agents-on-windowsagentarena-50-steps
Agents
Computer Use Agents
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Show papers without code
Paper
Code
Success Rate
↕
ModelName
ReleaseDate
↕
UI-Mate: Advancing Open-Weight Foundation GUI Agents with In-Context Demonstrations
✓ Link
66.2
UI-Mate-27B
2026-08-16
OS-Helix: Unifying GUI Interaction and Code Execution in Native Computer-Using Agents
✓ Link
63.87
OS-Helix-27B-RL
2026-10-15
OS-Symphony: A Holistic Framework for Robust and Generalist Computer-Using Agent
✓ Link
63.45
OS-Symphony w/ GPT-5
2026-01-12
The Unreasonable Effectiveness of Scaling Agents for Computer Use
✓ Link
54.1
Agent S3 w/ bBoN (N=3) w/ GPT-5 (50-step)
2025-10-02
UI-TARS-2 Technical Report: Advancing GUI Agent with Multi-Turn Reinforcement Learning
✓ Link
50.6
UI-TARS-2
2025-09-02
CoAct-1: Computer-using Multi-Agent System with Coding Actions
43.5
CoAct-1
2025-08-05
UI-TARS: Pioneering Automated GUI Interaction with Native Agents
✓ Link
42.1
UI-TARS-1.5 (early access)
2025-01-21
Recovering Policy-Induced Errors: Benchmarking and Trajectory Synthesis for Robust GUI Agents
✓ Link
39.1
RoTS-32B
2026-05-28
HybridCUA: Learning to Orchestrate GUI and CLI for Computer-Use Agents
✓ Link
36.0
HybridCUA-9B
2026-09-29
ToolCUA: Towards Optimal GUI-Tool Path Orchestration for Computer Use Agents
✓ Link
33.8
ToolCUA-8B
2026-05-12
Scaling Computer-Use Grounding via User Interface Decomposition and Synthesis
✓ Link
32.8
JEDI-7B w/ GPT-4o
2025-05-19
ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data
✓ Link
24.2
ScaleCUA-32B (50 steps)
2025-09-18
×