| Paper | Code | Success Rate | ModelName | ReleaseDate |
|---|---|---|---|---|
| SCALECUA: Scaling Computer Use Agents with Verifiable Task Synthesis and Efficient Online RL | ✓ Link | 68.7 | ScaleCUA-Qwen3.5-9B (extra data) | 2026-07-13 |
| HybridCUA: Learning to Orchestrate GUI and CLI for Computer-Use Agents | ✓ Link | 53.6 | HybridCUA-9B | 2026-09-29 |
| GTA1: GUI Test-time Scaling Agent | ✓ Link | 43.4 | GTA1-7B-2507 + o3 planner/judge | 2025-07-08 |
| Agent S2: A Compositional Generalist-Specialist Framework for Computer Use Agents | ✓ Link | 34.5 | Agent S2 w/ Claude-3.7-Sonnet | 2025-04-01 |
| UI-TARS: Pioneering Automated GUI Interaction with Native Agents | ✓ Link | 24.6 | UI-TARS-72B-DPO | 2025-01-21 |
| ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data | ✓ Link | 17.7 | ScaleCUA-32B | 2025-09-18 |
| LiteCUA: Computer as MCP Server for Computer-Use Agent on AIOS | ✓ Link | 14.66 | LiteCUA w/ GPT-4o | 2025-05-24 |