{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ESUXJ6WVEESE7FHQEEKXXP32M7","short_pith_number":"pith:ESUXJ6WV","schema_version":"1.0","canonical_sha256":"24a974fad521244f94f021157bbf7a67d8eb2684ca3ce221a45357400ac0c35f","source":{"kind":"arxiv","id":"2508.09124","version":1},"attestation_state":"computed","paper":{"title":"OdysseyBench: Evaluating LLM Agents on Long-Horizon Complex Office Application Workflows","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daniel Madrigal Diaz, Dongge Han, Jin Xu, Saravan Rajmohan, Victor R\\\"uhle, Weixuan Wang","submitted_at":"2025-08-12T17:53:03Z","abstract_excerpt":"Autonomous agents powered by large language models (LLMs) are increasingly deployed in real-world applications requiring complex, long-horizon workflows. However, existing benchmarks predominantly focus on atomic tasks that are self-contained and independent, failing to capture the long-term contextual dependencies and multi-interaction coordination required in realistic scenarios. To address this gap, we introduce OdysseyBench, a comprehensive benchmark for evaluating LLM agents on long-horizon workflows across diverse office applications including Word, Excel, PDF, Email, and Calendar. Our b"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2508.09124","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-08-12T17:53:03Z","cross_cats_sorted":[],"title_canon_sha256":"172db8c838cc1e1fced8c3b16ea61c8f9e2b4080c0e922d98b4be018b489e54a","abstract_canon_sha256":"602b046a1e89d1f8f77aa61a0f2913508cfb6086b34220c450ca6c160d1f9b55"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:39.972024Z","signature_b64":"Rr/lIsEJbiwCjTA3EYdQVptBTv3R9LUT+i9Hgn0BnnfBu7nxOMMO94THfOISwlyPRAVDVhyZ8vZPODumdZHtBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"24a974fad521244f94f021157bbf7a67d8eb2684ca3ce221a45357400ac0c35f","last_reissued_at":"2026-07-05T11:52:39.971539Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:39.971539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OdysseyBench: Evaluating LLM Agents on Long-Horizon Complex Office Application Workflows","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daniel Madrigal Diaz, Dongge Han, Jin Xu, Saravan Rajmohan, Victor R\\\"uhle, Weixuan Wang","submitted_at":"2025-08-12T17:53:03Z","abstract_excerpt":"Autonomous agents powered by large language models (LLMs) are increasingly deployed in real-world applications requiring complex, long-horizon workflows. However, existing benchmarks predominantly focus on atomic tasks that are self-contained and independent, failing to capture the long-term contextual dependencies and multi-interaction coordination required in realistic scenarios. To address this gap, we introduce OdysseyBench, a comprehensive benchmark for evaluating LLM agents on long-horizon workflows across diverse office applications including Word, Excel, PDF, Email, and Calendar. Our b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.09124","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.09124/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2508.09124","created_at":"2026-07-05T11:52:39.971599+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.09124v1","created_at":"2026-07-05T11:52:39.971599+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.09124","created_at":"2026-07-05T11:52:39.971599+00:00"},{"alias_kind":"pith_short_12","alias_value":"ESUXJ6WVEESE","created_at":"2026-07-05T11:52:39.971599+00:00"},{"alias_kind":"pith_short_16","alias_value":"ESUXJ6WVEESE7FHQ","created_at":"2026-07-05T11:52:39.971599+00:00"},{"alias_kind":"pith_short_8","alias_value":"ESUXJ6WV","created_at":"2026-07-05T11:52:39.971599+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":16,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.06008","citing_title":"PolyWorkBench: Benchmarking Multilingual Long-Horizon LLM Agents","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2607.06008","citing_title":"PolyWorkBench: Benchmarking Multilingual Long-Horizon LLM Agents","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2606.26346","citing_title":"How Do Tool-Augmented LLM Agents Perform on Real-World Energy Analytics Tasks?","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21654","citing_title":"ChainWorld: Composing Long-Horizon Desktop Workloads from Atomic OSWorld Tasks","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01245","citing_title":"Office Comprehension Benchmark","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18543","citing_title":"CEO-Bench: Can Agents Play the Long Game?","ref_index":91,"is_internal_anchor":false},{"citing_arxiv_id":"2606.14397","citing_title":"Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08340","citing_title":"Benchmarking Open-Ended Multi-Agent Coordination in Language Agents","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05342","citing_title":"SentinelBench: A Benchmark for Long-Running Monitoring Agents","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15230","citing_title":"EnergyAgentBench: Benchmarking LLM Agents on Live Energy Infrastructure Data","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2512.13168","citing_title":"Finch: Benchmarking Finance & Accounting across Spreadsheet-Centric Enterprise Workflows","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02522","citing_title":"Opal: Private Memory for Personal AI","ref_index":239,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10912","citing_title":"WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23455","citing_title":"CUJBench: Benchmarking LLM-Agent on Cross-Modal Failure Diagnosis from Browser to Backend","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07339","citing_title":"Tools as Continuous Flow for Evolving Agentic Reasoning","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15715","citing_title":"GTA-2: Benchmarking General Tool Agents from Atomic Tool-Use to Open-Ended Workflows","ref_index":18,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7","json":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7.json","graph_json":"https://pith.science/api/pith-number/ESUXJ6WVEESE7FHQEEKXXP32M7/graph.json","events_json":"https://pith.science/api/pith-number/ESUXJ6WVEESE7FHQEEKXXP32M7/events.json","paper":"https://pith.science/paper/ESUXJ6WV"},"agent_actions":{"view_html":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7","download_json":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7.json","view_paper":"https://pith.science/paper/ESUXJ6WV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.09124&json=true","fetch_graph":"https://pith.science/api/pith-number/ESUXJ6WVEESE7FHQEEKXXP32M7/graph.json","fetch_events":"https://pith.science/api/pith-number/ESUXJ6WVEESE7FHQEEKXXP32M7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7/action/storage_attestation","attest_author":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7/action/author_attestation","sign_citation":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7/action/citation_signature","submit_replication":"https://pith.science/pith/ESUXJ6WVEESE7FHQEEKXXP32M7/action/replication_record"}},"created_at":"2026-07-05T11:52:39.971599+00:00","updated_at":"2026-07-05T11:52:39.971599+00:00"}