{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QLENAFDICCZ545TFP3QYRSIITN","short_pith_number":"pith:QLENAFDI","schema_version":"1.0","canonical_sha256":"82c8d0146810b3de76657ee188c9089b516ea76ecc0b0d7a6bff464835068760","source":{"kind":"arxiv","id":"2608.02642","version":1},"attestation_state":"computed","paper":{"title":"MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"physics.chem-ph","authors_text":"Eden James Gage, Jonathan David Colburn, Kevin Guan, Kyle Vaccaro, Linda Xi Phan, Minjoon Seo, Nithishwer Mouroug Anand, Philip C. Biggin, Wei-Tse Hsu","submitted_at":"2026-07-31T13:48:07Z","abstract_excerpt":"Accelerating scientific discovery is among the most consequential applications of AI, and computational biomolecular simulation stands out as a particularly promising target within this broader effort. Coding agents promise to automate significant portions of this workflow, yet their reliability on realistic molecular dynamics (MD) tasks remains poorly characterized. To address this issue, we introduce MDArena, a benchmark of 50 containerized tasks drawn from active biomolecular simulation projects, spanning 29 molecular systems and 14 broad research protocols, including trajectory analysis, c"},"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":"2608.02642","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"physics.chem-ph","submitted_at":"2026-07-31T13:48:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3b0f79dc225b63a85015e65e21fbcb24c05fed81f8b4d1760e673aa4c96a5c04","abstract_canon_sha256":"2f9c2308d5abd25cf8aac1904a93ca710431bde21c8707ceffcbe953a852f381"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T00:38:06.443219Z","signature_b64":"BAHkTO3ldEdvOYhPp5iwgzpX0VAtTwrLvFbh+rpgMIn8O0C/WpB5Kz4lrEFJQ/v0VGelR2x71JwSMg6e3PEwCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"82c8d0146810b3de76657ee188c9089b516ea76ecc0b0d7a6bff464835068760","last_reissued_at":"2026-08-05T00:38:06.440874Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T00:38:06.440874Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"physics.chem-ph","authors_text":"Eden James Gage, Jonathan David Colburn, Kevin Guan, Kyle Vaccaro, Linda Xi Phan, Minjoon Seo, Nithishwer Mouroug Anand, Philip C. Biggin, Wei-Tse Hsu","submitted_at":"2026-07-31T13:48:07Z","abstract_excerpt":"Accelerating scientific discovery is among the most consequential applications of AI, and computational biomolecular simulation stands out as a particularly promising target within this broader effort. Coding agents promise to automate significant portions of this workflow, yet their reliability on realistic molecular dynamics (MD) tasks remains poorly characterized. To address this issue, we introduce MDArena, a benchmark of 50 containerized tasks drawn from active biomolecular simulation projects, spanning 29 molecular systems and 14 broad research protocols, including trajectory analysis, c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.02642","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/2608.02642/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":"2608.02642","created_at":"2026-08-05T00:38:06.441744+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.02642v1","created_at":"2026-08-05T00:38:06.441744+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.02642","created_at":"2026-08-05T00:38:06.441744+00:00"},{"alias_kind":"pith_short_12","alias_value":"QLENAFDICCZ5","created_at":"2026-08-05T00:38:06.441744+00:00"},{"alias_kind":"pith_short_16","alias_value":"QLENAFDICCZ545TF","created_at":"2026-08-05T00:38:06.441744+00:00"},{"alias_kind":"pith_short_8","alias_value":"QLENAFDI","created_at":"2026-08-05T00:38:06.441744+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/QLENAFDICCZ545TFP3QYRSIITN","json":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN.json","graph_json":"https://pith.science/api/pith-number/QLENAFDICCZ545TFP3QYRSIITN/graph.json","events_json":"https://pith.science/api/pith-number/QLENAFDICCZ545TFP3QYRSIITN/events.json","paper":"https://pith.science/paper/QLENAFDI"},"agent_actions":{"view_html":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN","download_json":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN.json","view_paper":"https://pith.science/paper/QLENAFDI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.02642&json=true","fetch_graph":"https://pith.science/api/pith-number/QLENAFDICCZ545TFP3QYRSIITN/graph.json","fetch_events":"https://pith.science/api/pith-number/QLENAFDICCZ545TFP3QYRSIITN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN/action/storage_attestation","attest_author":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN/action/author_attestation","sign_citation":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN/action/citation_signature","submit_replication":"https://pith.science/pith/QLENAFDICCZ545TFP3QYRSIITN/action/replication_record"}},"created_at":"2026-08-05T00:38:06.441744+00:00","updated_at":"2026-08-05T00:38:06.441744+00:00"}