Last updated Aug 07, 2026
OSWorld-Verified Leaderboard
The official leaderboard's latest evaluation data, as reported by our Scout
Foundation E2E GUI selected
General and specialized models without extra accessibility-tree or coding-action channels.
| Model | Institution | Approach type | Max steps | Success rate | Std. dev. | Runs | Date | A11y tree | Coding action | Multi-rollout |
|---|---|---|---|---|---|---|---|---|---|---|
| claude-fable-5[1m] | Anthropic | General model | 100 | 86.0% | 0.0 | 1 | Aug 01, 2026 | No | No | No |
| claude-opus-5[1m] | Anthropic | General model | 100 | 83.4% | 0.0 | 1 | Aug 01, 2026 | No | No | No |
| Muse Spark 1.1 | Meta Superintelligence Labs | General model | 100 | 80.7% | 0.0 | 1 | Jul 09, 2026 | No | No | No |
| Holo3-35B-A3B | H Company | Specialized model | 100 | 80.4% | 2.2 | 2 | Apr 20, 2026 | No | No | No |
| MiniMax M3 | MiniMax | General model | 100 | 75.2% | 0.0 | 1 | Jun 08, 2026 | No | No | No |
| Qwen 3.7 Plus | Qwen Team, Alibaba Group | General model | 100 | 73.3% | 0.0 | 1 | May 25, 2026 | No | No | No |
| Kimi K2.6 | Moonshot AI | General model | 100 | 73.1% | 0.0 | 1 | Apr 20, 2026 | No | No | No |
| claude-sonnet-4-6 | Anthropic | General model | 100 | 72.1% | 0.0 | 1 | Mar 08, 2026 | No | No | No |
| Kimi K2.5 | Moonshot AI | General model | 100 | 63.3% | 0.0 | 1 | Jan 31, 2026 | No | No | No |
| claude-sonnet-4-5-20250929 | Anthropic | General model | 100 | 62.9% | 0.0 | 1 | Oct 31, 2025 | No | No | No |
| Seed-1.8 | ByteDance Seed | General model | 100 | 61.9% | 0.0 | 1 | Dec 18, 2025 | No | No | No |
| claude-sonnet-4-5-20250929 | Anthropic | General model | 50 | 58.1% | 0.0 | 1 | Oct 31, 2025 | No | No | No |
| EvoCUA-20260105 | Meituan LongCat Team | General model | 50 | 56.7% | 0.0 | 1 | Jan 06, 2026 | No | No | No |
| GUI-Owl-1.5 32B | Tongyi Lab, Alibaba Group | Specialized model | 50 | 55.4% | 0.0 | 1 | Mar 27, 2026 | No | No | No |
| DeepMiner-Mano-72B | Mininglamp Technology | Specialized model | 100 | 53.9% | 0.0 | 1 | Oct 31, 2025 | No | No | No |
| UI-TARS-2-2509 | ByteDance Seed | Specialized model | 100 | 53.1% | 0.0 | 1 | Oct 14, 2025 | No | No | No |
| EvoCUA | Meituan LongCat Team | General model | 50 | 50.3% | 0.0 | 1 | Dec 24, 2025 | No | No | No |
| EvoCUA-8B-20260105 | Meituan LongCat Team | General model | 50 | 46.1% | 0.0 | 1 | Jan 13, 2026 | No | No | No |
| opencua-72b-preview | The University of Hong Kong & Moonshot AI | Specialized model | 100 | 45.0% | 0.9 | 3 | Oct 01, 2025 | No | No | No |
| opencua-72b-preview | The University of Hong Kong & Moonshot AI | Specialized model | 50 | 44.9% | 0.0 | 1 | Oct 01, 2025 | No | No | No |
| claude-4-sonnet-20250514 | Anthropic | General model | 50 | 43.9% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| claude-sonnet-4-5-20250929 | Anthropic | General model | 15 | 42.9% | 0.0 | 1 | Oct 31, 2025 | No | No | No |
| UI-TARS-250705 | ByteDance Seed | General model | 100 | 41.8% | 0.0 | 1 | Sep 16, 2025 | No | No | No |
| qwen3-vl-flash-2025-10-25 | Qwen Team, Alibaba Group | General model | 100 | 41.6% | 0.0 | 1 | Oct 23, 2025 | No | No | No |
| claude-4-sonnet-20250514 | Anthropic | General model | 100 | 41.4% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| DART-GUI-7B-0924 | BIGAI & BIT & DataCanvas | Specialized model | 30 | 40.5% | 0.0 | 1 | Jan 23, 2026 | No | No | No |
| DeepMiner-Mano-7B | Mininglamp Technology | Specialized model | 100 | 40.1% | 0.0 | 1 | Sep 18, 2025 | No | No | No |
| doubao-1-5-thinking-vision-pro-250717 | ByteDance Seed | General model | 100 | 40.0% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-72b-preview | The University of Hong Kong & Moonshot AI | Specialized model | 15 | 39.0% | 0.0 | 1 | Oct 01, 2025 | No | No | No |
| claude-3-7-sonnet-20250219 | Anthropic | General model | 50 | 35.8% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| claude-3-7-sonnet-20250219 | Anthropic | General model | 100 | 35.6% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-32b | The University of Hong Kong & Moonshot AI | Specialized model | 100 | 34.8% | 0.8 | 3 | Jul 28, 2025 | No | No | No |
| opencua-32b | The University of Hong Kong & Moonshot AI | Specialized model | 50 | 34.1% | 0.7 | 3 | Jul 28, 2025 | No | No | No |
| doubao-1-5-thinking-vision-pro-250428 | ByteDance Seed | General model | 100 | 33.8% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| gui-owl-7b | Tongyi Lab, Alibaba Group | Specialized model | 15 | 32.1% | 0.0 | 1 | Sep 03, 2025 | No | No | No |
| doubao-1-5-thinking-vision-pro-250717 | ByteDance Seed | General model | 15 | 31.9% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| computer-use-preview | OpenAI | Specialized model | 50 | 31.3% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| claude-4-sonnet-20250514 | Anthropic | General model | 15 | 31.2% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| computer-use-preview | OpenAI | Specialized model | 100 | 30.5% | 0.9 | 2 | Jul 28, 2025 | No | No | No |
| TianXi-Action-7B | Lenovo | Specialized model | 50 | 29.8% | 0.6 | 2 | Aug 11, 2025 | No | No | No |
| opencua-32b | The University of Hong Kong & Moonshot AI | Specialized model | 15 | 29.7% | 1.1 | 3 | Jul 28, 2025 | No | No | No |
| opencua-7b | The University of Hong Kong & Moonshot AI | Specialized model | 50 | 28.2% | 0.5 | 3 | Jul 28, 2025 | No | No | No |
| doubao-1-5-thinking-vision-pro-250428 | ByteDance Seed | General model | 15 | 27.8% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| uitars-1.5-7b | ByteDance Seed & Tsinghua University | Specialized model | 100 | 27.4% | 2.2 | 2 | Jul 28, 2025 | No | No | No |
| uitars-1.5-7b | ByteDance Seed & Tsinghua University | Specialized model | 50 | 27.3% | 2.1 | 2 | Jul 28, 2025 | No | No | No |
| claude-3-7-sonnet-20250219 | Anthropic | General model | 15 | 27.1% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| uitars-72b-dpo | ByteDance Seed & Tsinghua University | Specialized model | 100 | 27.1% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-7b | The University of Hong Kong & Moonshot AI | Specialized model | 100 | 26.6% | 0.4 | 3 | Jul 28, 2025 | No | No | No |
| computer-use-preview | OpenAI | Specialized model | 15 | 26.0% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| uitars-72b-dpo | ByteDance Seed & Tsinghua University | Specialized model | 50 | 25.8% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| uitars-1.5-7b | ByteDance Seed & Tsinghua University | Specialized model | 15 | 24.5% | 1.2 | 2 | Jul 28, 2025 | No | No | No |
| opencua-7b | The University of Hong Kong & Moonshot AI | Specialized model | 15 | 24.3% | 1.4 | 3 | Jul 28, 2025 | No | No | No |
| uitars-72b-dpo | ByteDance Seed & Tsinghua University | Specialized model | 15 | 24.0% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-qwen2-7b | The University of Hong Kong & Moonshot AI | Specialized model | 100 | 23.1% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| o3 | OpenAI | General model | 100 | 23.0% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-qwen2-7b | The University of Hong Kong & Moonshot AI | Specialized model | 50 | 20.6% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-a3b | The University of Hong Kong & Moonshot AI | Specialized model | 50 | 19.9% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-qwen2-7b | The University of Hong Kong & Moonshot AI | Specialized model | 15 | 19.9% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-a3b | The University of Hong Kong & Moonshot AI | Specialized model | 100 | 17.7% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| o3 | OpenAI | General model | 50 | 17.2% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| opencua-a3b | The University of Hong Kong & Moonshot AI | Specialized model | 15 | 16.9% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| kimi-vl-a3b | Moonshot AI | General model | 100 | 10.3% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| kimi-vl-a3b | Moonshot AI | General model | 15 | 9.7% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| o3 | OpenAI | General model | 15 | 9.1% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| qwen2.5-vl-72b-instruct | Alibaba Cloud, Qwen Team | General model | 100 | 5.0% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| qwen2.5-vl-72b-instruct | Alibaba Cloud, Qwen Team | General model | 15 | 4.4% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| qwen2.5-vl-32b-instruct | Alibaba Cloud, Qwen Team | General model | 100 | 3.9% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
| qwen2.5-vl-32b-instruct | Alibaba Cloud, Qwen Team | General model | 15 | 3.0% | 0.0 | 1 | Jul 28, 2025 | No | No | No |
Frequently asked questions
Resources
Paper (arXiv 2404.07972)Code (GitHub · XLANG)Verified trajectories (Hugging Face)Official leaderboard
Citation
@misc{OSWorld,
title={OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments},
author={Tianbao Xie and Danyang Zhang and Jixuan Chen and Xiaochuan Li and Siheng Zhao and Ruisheng Cao and Toh Jing Hua and Zhoujun Cheng and Dongchan Shin and Fangyu Lei and Yitao Liu and Yiheng Xu and Shuyan Zhou and Silvio Savarese and Caiming Xiong and Victor Zhong and Tao Yu},
year={2024},
eprint={2404.07972},
archivePrefix={arXiv},
primaryClass={cs.AI},
}