{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BJQH5Z2KADOFE6JGH6FQGUFJCK","short_pith_number":"pith:BJQH5Z2K","schema_version":"1.0","canonical_sha256":"0a607ee74a00dc5279263f8b0350a912a65dfc23ba3d66b24bbc6d41c033bb3e","source":{"kind":"arxiv","id":"2606.07682","version":1},"attestation_state":"computed","paper":{"title":"SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Adnan El Assadi, Albert Liu, Christopher Settles, Daniel Wang, Derek Chen, Erik Quintanilla, Fenil Faldu, Ishan Gupta, Ivan Bercovich, Jesse Hu, Jiankai Sun, Joan Cabezas, Luyang Kong, Marek \\v{S}uppa, Neel Harsola, Nevasini Sasikumar, Omkaar Mukund Kamath, Pramod Srinivasan, Pranav Raja, Prannay Hebbar, Pratyush Shukla, Rishi Desai, Roey Ben Chaim, Steven Dillmann, Xiangyi Li, Yiyuan Li","submitted_at":"2026-06-05T00:39:44Z","abstract_excerpt":"AI agents are increasingly expected to complete long-horizon workflows that require sustained progress over hours, millions of tokens, and complex environments. Yet current agent benchmarks largely evaluate short-form tasks, such as single pull requests, small tickets, or 5-10 minute exercises, limiting our ability to measure agents' capabilities in planning, long-context understanding, and memory use. We introduce SWE-Marathon, a benchmark of 20 long-horizon tasks spanning software engineering and adjacent technical domains. Each task consists of a unique executable environment, a human-writt"},"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":"2606.07682","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-06-05T00:39:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"64ea08d3ae672f0f30406644490f64b0d58707833f58949e7080616cecb784e0","abstract_canon_sha256":"2b409f0d84bb8d60f3c495d2d0a78a06c81eade78055c2cc2093e3d6eaaef993"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T00:04:46.730435Z","signature_b64":"0xqA636TGic46/yWbz6sMkhuO5yPRFM4J8v6ZJ5Yibnn5z01XVjsX9tHCfKtjiDXZWEnR1ADRj6Selz2hp/7Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0a607ee74a00dc5279263f8b0350a912a65dfc23ba3d66b24bbc6d41c033bb3e","last_reissued_at":"2026-06-09T00:04:46.729996Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T00:04:46.729996Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SWE-Marathon: Can Agents Autonomously Complete Ultra-Long-Horizon Software Work?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Adnan El Assadi, Albert Liu, Christopher Settles, Daniel Wang, Derek Chen, Erik Quintanilla, Fenil Faldu, Ishan Gupta, Ivan Bercovich, Jesse Hu, Jiankai Sun, Joan Cabezas, Luyang Kong, Marek \\v{S}uppa, Neel Harsola, Nevasini Sasikumar, Omkaar Mukund Kamath, Pramod Srinivasan, Pranav Raja, Prannay Hebbar, Pratyush Shukla, Rishi Desai, Roey Ben Chaim, Steven Dillmann, Xiangyi Li, Yiyuan Li","submitted_at":"2026-06-05T00:39:44Z","abstract_excerpt":"AI agents are increasingly expected to complete long-horizon workflows that require sustained progress over hours, millions of tokens, and complex environments. Yet current agent benchmarks largely evaluate short-form tasks, such as single pull requests, small tickets, or 5-10 minute exercises, limiting our ability to measure agents' capabilities in planning, long-context understanding, and memory use. We introduce SWE-Marathon, a benchmark of 20 long-horizon tasks spanning software engineering and adjacent technical domains. Each task consists of a unique executable environment, a human-writt"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07682","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/2606.07682/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":"2606.07682","created_at":"2026-06-09T00:04:46.730067+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.07682v1","created_at":"2026-06-09T00:04:46.730067+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07682","created_at":"2026-06-09T00:04:46.730067+00:00"},{"alias_kind":"pith_short_12","alias_value":"BJQH5Z2KADOF","created_at":"2026-06-09T00:04:46.730067+00:00"},{"alias_kind":"pith_short_16","alias_value":"BJQH5Z2KADOFE6JG","created_at":"2026-06-09T00:04:46.730067+00:00"},{"alias_kind":"pith_short_8","alias_value":"BJQH5Z2K","created_at":"2026-06-09T00:04:46.730067+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK","json":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK.json","graph_json":"https://pith.science/api/pith-number/BJQH5Z2KADOFE6JGH6FQGUFJCK/graph.json","events_json":"https://pith.science/api/pith-number/BJQH5Z2KADOFE6JGH6FQGUFJCK/events.json","paper":"https://pith.science/paper/BJQH5Z2K"},"agent_actions":{"view_html":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK","download_json":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK.json","view_paper":"https://pith.science/paper/BJQH5Z2K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.07682&json=true","fetch_graph":"https://pith.science/api/pith-number/BJQH5Z2KADOFE6JGH6FQGUFJCK/graph.json","fetch_events":"https://pith.science/api/pith-number/BJQH5Z2KADOFE6JGH6FQGUFJCK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK/action/storage_attestation","attest_author":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK/action/author_attestation","sign_citation":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK/action/citation_signature","submit_replication":"https://pith.science/pith/BJQH5Z2KADOFE6JGH6FQGUFJCK/action/replication_record"}},"created_at":"2026-06-09T00:04:46.730067+00:00","updated_at":"2026-06-09T00:04:46.730067+00:00"}