AN Interesting Verifiable Benchmark for Autonomous Agents in Complex Document Operations.<br>our code: <a href=\"https://github.com/icip-cas/DocOps\" rel=\"nofollow\">https://github.com/icip-cas/DocOps</a><br>our page: <a href=\"https://docopsbench.github.io\" rel=\"nofollow\">https://docopsbench.github.io</a><br>our arxiv: <a href=\"https://arxiv.org/abs/2607.19865\" rel=\"nofollow\">https://arxiv.org/abs/2607.19865</a></p>\n","updatedAt":"2026-07-23T08:54:37.714Z","author":{"_id":"6a61815980280a26ac9d3941","avatarUrl":"/avatars/267590e5a6b214bb018f28f4e6cff101.svg","fullname":"Jiazhen Jiang","name":"jiangxiaohuan","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5463017821311951},"editors":["jiangxiaohuan"],"editorAvatarUrls":["/avatars/267590e5a6b214bb018f28f4e6cff101.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2607.19865","authors":[{"_id":"6a618294c3792c34f5a040de","user":{"_id":"6a61815980280a26ac9d3941","avatarUrl":"/avatars/267590e5a6b214bb018f28f4e6cff101.svg","isPro":false,"fullname":"Jiazhen Jiang","user":"jiangxiaohuan","type":"user","name":"jiangxiaohuan"},"name":"Jiazhen Jiang","status":"claimed_verified","statusLastChangedAt":"2026-07-23T08:45:04.640Z","hidden":false},{"_id":"6a618294c3792c34f5a040df","name":"Boxi Cao","hidden":false},{"_id":"6a618294c3792c34f5a040e0","name":"Lingyong Yan","hidden":false},{"_id":"6a618294c3792c34f5a040e1","name":"Yaojie Lu","hidden":false},{"_id":"6a618294c3792c34f5a040e2","name":"Hongyu Lin","hidden":false},{"_id":"6a618294c3792c34f5a040e3","name":"Shuaiqiang Wang","hidden":false},{"_id":"6a618294c3792c34f5a040e4","name":"Dawei Yin","hidden":false},{"_id":"6a618294c3792c34f5a040e5","name":"Xianpei Han","hidden":false},{"_id":"6a618294c3792c34f5a040e6","name":"Le Sun","hidden":false}],"publishedAt":"2026-07-22T00:00:00.000Z","submittedOnDailyAt":"2026-07-23T00:00:00.000Z","title":"DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations","submittedOnDailyBy":{"_id":"6a61815980280a26ac9d3941","avatarUrl":"/avatars/267590e5a6b214bb018f28f4e6cff101.svg","isPro":false,"fullname":"Jiazhen Jiang","user":"jiangxiaohuan","type":"user","name":"jiangxiaohuan"},"summary":"As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, revealing that even the most advanced frontier configurations still exhibit profound limitations when handling highly coupled, long-range tasks. Furthermore, a fine-grained analysis of existing agents' manipulation behaviors uncovers 3 key failure modes: long-term state tracking collapse, shallow semantic verification, and destructive editing of structural metadata. Ultimately, our work exposes the capability boundaries of agents in maintaining global document consistency, shedding light on the future design of robust, non-destructive agents for complex digital ecosystems.","upvotes":3,"discussionId":"6a618294c3792c34f5a040e7","projectPage":"https://docopsbench.github.io","githubRepo":"https://github.com/icip-cas/DocOps","githubRepoAddedBy":"user","githubStars":1},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6a61815980280a26ac9d3941","avatarUrl":"/avatars/267590e5a6b214bb018f28f4e6cff101.svg","isPro":false,"fullname":"Jiazhen Jiang","user":"jiangxiaohuan","type":"user"},{"_id":"64cde36344d373d706f2e7c2","avatarUrl":"/avatars/a4089c30642fd5f0d5e49245e6ab95bf.svg","isPro":false,"fullname":"王蓬勃","user":"roninpb","type":"user"},{"_id":"5fc052241160c47d1d438556","avatarUrl":"/avatars/f508707fc92f1b42f7897b12b727754c.svg","isPro":false,"fullname":"Boxi Cao","user":"Bowieee","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"query":{}}">
DocOps: A Verifiable Benchmark for Autonomous Agents in Complex Document Operations
Abstract
As autonomous agents rapidly evolve, their ability to reliably manipulate ubiquitous digital documents has become critical for enabling general-purpose AI assistants and automating complex workspace workflows. In this paper, we introduce DocOps, a deterministically verifiable evaluation framework underpinned by a hierarchical taxonomy that deconstructs document operations inspired by real-world practices into atomic dimensions and escalating workflow complexities. Based on DocOps, we systematically evaluate representative closed- and open-source models across various agentic harnesses, revealing that even the most advanced frontier configurations still exhibit profound limitations when handling highly coupled, long-range tasks. Furthermore, a fine-grained analysis of existing agents' manipulation behaviors uncovers 3 key failure modes: long-term state tracking collapse, shallow semantic verification, and destructive editing of structural metadata. Ultimately, our work exposes the capability boundaries of agents in maintaining global document consistency, shedding light on the future design of robust, non-destructive agents for complex digital ecosystems.
Community
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2607.19865 in a model README.md to link it from this page.
Cite arxiv.org/abs/2607.19865 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2607.19865 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.