Current benchmarks for computer-use agents evaluate models in impersonal environments. This leaves a gap between evaluation and deployment where personal assistants are expected to work across a user's whole digital life, including their context, historical data, and logged-in accounts. This gap is widest on web tasks, where live web evaluations cannot exercise sites that require logging in or personal information, the kind of site a real personal assistant has to drive. We introduce MyPCBench, which tests computer-use agents as personal assistants on a Linux desktop populated with 17 simulated real-world web applications and a full desktop stack, all seeded for one canonical persona, Michael Scott from The Office. We define 184 tasks in this environment, each inspired by a real request drawn from the OpenClaw community, and benchmark six closed and open-weight models with a uniform computer+bash tool surface. We find that the best model, Claude Opus 4.6, fully solves 55.4% of the tasks, the only model above 50%. Model failures cluster on tasks that span many applications and on long trajectories, where personalization stresses an assistant the most. We release the environment, task set, and agent harness at mypcbench.com</p>\n","updatedAt":"2026-06-18T13:02:47.832Z","author":{"_id":"664aebe829eadb3ab4e4ca3f","avatarUrl":"/avatars/548d5656e082c8959ac78b883f0805af.svg","fullname":"Lawrence Jang","name":"ljang0","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.9245612621307373},"editors":["ljang0"],"editorAvatarUrls":["/avatars/548d5656e082c8959ac78b883f0805af.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.16748","authors":[{"_id":"6a319104bc818ff14e453bbf","name":"Lawrence Keunho Jang","hidden":false},{"_id":"6a319104bc818ff14e453bc0","name":"Andrew Keunwoo Jang","hidden":false},{"_id":"6a319104bc818ff14e453bc1","name":"Jing Yu Koh","hidden":false},{"_id":"6a319104bc818ff14e453bc2","name":"Ruslan Salakhutdinov","hidden":false}],"publishedAt":"2026-06-15T00:00:00.000Z","submittedOnDailyAt":"2026-06-18T00:00:00.000Z","title":"MyPCBench: A Benchmark for Personally Intelligent Computer-Use Agents","submittedOnDailyBy":{"_id":"664aebe829eadb3ab4e4ca3f","avatarUrl":"/avatars/548d5656e082c8959ac78b883f0805af.svg","isPro":false,"fullname":"Lawrence Jang","user":"ljang0","type":"user","name":"ljang0"},"summary":"Current benchmarks for computer-use agents evaluate models in impersonal environments. This leaves a gap between evaluation and deployment where personal assistants are expected to work across a user's whole digital life, including their context, historical data, and logged-in accounts. This gap is widest on web tasks, where live web evaluations cannot exercise sites that require logging in or personal information, the kind of site a real personal assistant has to drive. We introduce MyPCBench, which tests computer-use agents as personal assistants on a Linux desktop populated with 17 simulated real-world web applications and a full desktop stack, all seeded for one canonical persona, Michael Scott from The Office. We define 184 tasks in this environment, each inspired by a real request drawn from the OpenClaw community, and benchmark six closed and open-weight models with a uniform computer+bash tool surface. We find that the best model, Claude Opus 4.6, fully solves 55.4\\% of the tasks, the only model above 50\\%. Model failures cluster on tasks that span many applications and on long trajectories, where personalization stresses an assistant the most. We release the environment, task set, and agent harness at https://mypcbench.com.","upvotes":1,"discussionId":"6a319104bc818ff14e453bc3","projectPage":"https://mypcbench.com/","githubRepo":"https://github.com/ljang0/MyPCBench","githubRepoAddedBy":"user","ai_summary":"MyPCBench evaluates computer-use agents as personal assistants in a simulated Linux desktop environment with real-world web applications, revealing that Claude Opus 4.6 achieves the highest task completion rate of 55.4% while struggles with multi-application tasks and long trajectories.","ai_keywords":["computer-use agents","personal assistants","web tasks","live web evaluations","digital life","contextual awareness","web applications","task completion","model performance","agent harness"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":3,"organization":{"_id":"691d9a1012cc4d473e1c862f","name":"CarnegieMellonU","fullname":"Carnegie Mellon University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/68e396f2b5bb631e9b2fac9a/6I146aJvxxlRCEbYFFAeQ.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"664aebe829eadb3ab4e4ca3f","avatarUrl":"/avatars/548d5656e082c8959ac78b883f0805af.svg","isPro":false,"fullname":"Lawrence Jang","user":"ljang0","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"691d9a1012cc4d473e1c862f","name":"CarnegieMellonU","fullname":"Carnegie Mellon University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/68e396f2b5bb631e9b2fac9a/6I146aJvxxlRCEbYFFAeQ.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2606/2606.16748.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2606.16748 in a model README.md to link it from this page.

Cite arxiv.org/abs/2606.16748 in a dataset README.md to link it from this page.

Cite arxiv.org/abs/2606.16748 in a Space README.md to link it from this page.