Yutori

Last updated Aug 07, 2026

OSWorld-Verified Leaderboard

The official leaderboard's latest evaluation data, as reported by our Scout

Markdown for agents

Foundation E2E GUI selected

General and specialized models without extra accessibility-tree or coding-action channels.

ModelInstitutionApproach typeMax stepsSuccess rateStd. dev.RunsDateA11y treeCoding actionMulti-rollout
claude-fable-5[1m]AnthropicGeneral model10086.0%0.01Aug 01, 2026NoNoNo
claude-opus-5[1m]AnthropicGeneral model10083.4%0.01Aug 01, 2026NoNoNo
Muse Spark 1.1Meta Superintelligence LabsGeneral model10080.7%0.01Jul 09, 2026NoNoNo
Holo3-35B-A3BH CompanySpecialized model10080.4%2.22Apr 20, 2026NoNoNo
MiniMax M3MiniMaxGeneral model10075.2%0.01Jun 08, 2026NoNoNo
Qwen 3.7 PlusQwen Team, Alibaba GroupGeneral model10073.3%0.01May 25, 2026NoNoNo
Kimi K2.6Moonshot AIGeneral model10073.1%0.01Apr 20, 2026NoNoNo
claude-sonnet-4-6AnthropicGeneral model10072.1%0.01Mar 08, 2026NoNoNo
Kimi K2.5Moonshot AIGeneral model10063.3%0.01Jan 31, 2026NoNoNo
claude-sonnet-4-5-20250929AnthropicGeneral model10062.9%0.01Oct 31, 2025NoNoNo
Seed-1.8ByteDance SeedGeneral model10061.9%0.01Dec 18, 2025NoNoNo
claude-sonnet-4-5-20250929AnthropicGeneral model5058.1%0.01Oct 31, 2025NoNoNo
EvoCUA-20260105Meituan LongCat TeamGeneral model5056.7%0.01Jan 06, 2026NoNoNo
GUI-Owl-1.5 32BTongyi Lab, Alibaba GroupSpecialized model5055.4%0.01Mar 27, 2026NoNoNo
DeepMiner-Mano-72BMininglamp TechnologySpecialized model10053.9%0.01Oct 31, 2025NoNoNo
UI-TARS-2-2509ByteDance SeedSpecialized model10053.1%0.01Oct 14, 2025NoNoNo
EvoCUAMeituan LongCat TeamGeneral model5050.3%0.01Dec 24, 2025NoNoNo
EvoCUA-8B-20260105Meituan LongCat TeamGeneral model5046.1%0.01Jan 13, 2026NoNoNo
opencua-72b-previewThe University of Hong Kong & Moonshot AISpecialized model10045.0%0.93Oct 01, 2025NoNoNo
opencua-72b-previewThe University of Hong Kong & Moonshot AISpecialized model5044.9%0.01Oct 01, 2025NoNoNo
claude-4-sonnet-20250514AnthropicGeneral model5043.9%0.01Jul 28, 2025NoNoNo
claude-sonnet-4-5-20250929AnthropicGeneral model1542.9%0.01Oct 31, 2025NoNoNo
UI-TARS-250705ByteDance SeedGeneral model10041.8%0.01Sep 16, 2025NoNoNo
qwen3-vl-flash-2025-10-25Qwen Team, Alibaba GroupGeneral model10041.6%0.01Oct 23, 2025NoNoNo
claude-4-sonnet-20250514AnthropicGeneral model10041.4%0.01Jul 28, 2025NoNoNo
DART-GUI-7B-0924BIGAI & BIT & DataCanvasSpecialized model3040.5%0.01Jan 23, 2026NoNoNo
DeepMiner-Mano-7BMininglamp TechnologySpecialized model10040.1%0.01Sep 18, 2025NoNoNo
doubao-1-5-thinking-vision-pro-250717ByteDance SeedGeneral model10040.0%0.01Jul 28, 2025NoNoNo
opencua-72b-previewThe University of Hong Kong & Moonshot AISpecialized model1539.0%0.01Oct 01, 2025NoNoNo
claude-3-7-sonnet-20250219AnthropicGeneral model5035.8%0.01Jul 28, 2025NoNoNo
claude-3-7-sonnet-20250219AnthropicGeneral model10035.6%0.01Jul 28, 2025NoNoNo
opencua-32bThe University of Hong Kong & Moonshot AISpecialized model10034.8%0.83Jul 28, 2025NoNoNo
opencua-32bThe University of Hong Kong & Moonshot AISpecialized model5034.1%0.73Jul 28, 2025NoNoNo
doubao-1-5-thinking-vision-pro-250428ByteDance SeedGeneral model10033.8%0.01Jul 28, 2025NoNoNo
gui-owl-7bTongyi Lab, Alibaba GroupSpecialized model1532.1%0.01Sep 03, 2025NoNoNo
doubao-1-5-thinking-vision-pro-250717ByteDance SeedGeneral model1531.9%0.01Jul 28, 2025NoNoNo
computer-use-previewOpenAISpecialized model5031.3%0.01Jul 28, 2025NoNoNo
claude-4-sonnet-20250514AnthropicGeneral model1531.2%0.01Jul 28, 2025NoNoNo
computer-use-previewOpenAISpecialized model10030.5%0.92Jul 28, 2025NoNoNo
TianXi-Action-7BLenovoSpecialized model5029.8%0.62Aug 11, 2025NoNoNo
opencua-32bThe University of Hong Kong & Moonshot AISpecialized model1529.7%1.13Jul 28, 2025NoNoNo
opencua-7bThe University of Hong Kong & Moonshot AISpecialized model5028.2%0.53Jul 28, 2025NoNoNo
doubao-1-5-thinking-vision-pro-250428ByteDance SeedGeneral model1527.8%0.01Jul 28, 2025NoNoNo
uitars-1.5-7bByteDance Seed & Tsinghua UniversitySpecialized model10027.4%2.22Jul 28, 2025NoNoNo
uitars-1.5-7bByteDance Seed & Tsinghua UniversitySpecialized model5027.3%2.12Jul 28, 2025NoNoNo
claude-3-7-sonnet-20250219AnthropicGeneral model1527.1%0.01Jul 28, 2025NoNoNo
uitars-72b-dpoByteDance Seed & Tsinghua UniversitySpecialized model10027.1%0.01Jul 28, 2025NoNoNo
opencua-7bThe University of Hong Kong & Moonshot AISpecialized model10026.6%0.43Jul 28, 2025NoNoNo
computer-use-previewOpenAISpecialized model1526.0%0.01Jul 28, 2025NoNoNo
uitars-72b-dpoByteDance Seed & Tsinghua UniversitySpecialized model5025.8%0.01Jul 28, 2025NoNoNo
uitars-1.5-7bByteDance Seed & Tsinghua UniversitySpecialized model1524.5%1.22Jul 28, 2025NoNoNo
opencua-7bThe University of Hong Kong & Moonshot AISpecialized model1524.3%1.43Jul 28, 2025NoNoNo
uitars-72b-dpoByteDance Seed & Tsinghua UniversitySpecialized model1524.0%0.01Jul 28, 2025NoNoNo
opencua-qwen2-7bThe University of Hong Kong & Moonshot AISpecialized model10023.1%0.01Jul 28, 2025NoNoNo
o3OpenAIGeneral model10023.0%0.01Jul 28, 2025NoNoNo
opencua-qwen2-7bThe University of Hong Kong & Moonshot AISpecialized model5020.6%0.01Jul 28, 2025NoNoNo
opencua-a3bThe University of Hong Kong & Moonshot AISpecialized model5019.9%0.01Jul 28, 2025NoNoNo
opencua-qwen2-7bThe University of Hong Kong & Moonshot AISpecialized model1519.9%0.01Jul 28, 2025NoNoNo
opencua-a3bThe University of Hong Kong & Moonshot AISpecialized model10017.7%0.01Jul 28, 2025NoNoNo
o3OpenAIGeneral model5017.2%0.01Jul 28, 2025NoNoNo
opencua-a3bThe University of Hong Kong & Moonshot AISpecialized model1516.9%0.01Jul 28, 2025NoNoNo
kimi-vl-a3bMoonshot AIGeneral model10010.3%0.01Jul 28, 2025NoNoNo
kimi-vl-a3bMoonshot AIGeneral model159.7%0.01Jul 28, 2025NoNoNo
o3OpenAIGeneral model159.1%0.01Jul 28, 2025NoNoNo
qwen2.5-vl-72b-instructAlibaba Cloud, Qwen TeamGeneral model1005.0%0.01Jul 28, 2025NoNoNo
qwen2.5-vl-72b-instructAlibaba Cloud, Qwen TeamGeneral model154.4%0.01Jul 28, 2025NoNoNo
qwen2.5-vl-32b-instructAlibaba Cloud, Qwen TeamGeneral model1003.9%0.01Jul 28, 2025NoNoNo
qwen2.5-vl-32b-instructAlibaba Cloud, Qwen TeamGeneral model153.0%0.01Jul 28, 2025NoNoNo

Frequently asked questions

Resources

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},
}