{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:EEGA3OZOOWWIFNCFEEB73HWRO2","short_pith_number":"pith:EEGA3OZO","schema_version":"1.0","canonical_sha256":"210c0dbb2e75ac82b4452103fd9ed1768d0a49e23c648efe7ede5af746a7f7c1","source":{"kind":"arxiv","id":"2607.15439","version":1},"attestation_state":"computed","paper":{"title":"Do Coding Agents Need Executable World Models, Simplification, and Verification to Solve ARC-AGI-3?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Sergey Rodionov","submitted_at":"2026-07-16T20:18:59Z","abstract_excerpt":"Our previous ARC-AGI-3 agent bundled executable world modeling, scheduled simplification, and exact replay verification, leaving unclear which idea accounted for its performance. We address this attribution question with four nested Codex-based agents: a textual baseline; a flexible-interface executable world model without replay verification; the same executable model with scheduled simplification; and a fixed-interface verification treatment that retains simplification and requires exact reproduction of recorded observations. The main study evaluates all four agents with gpt-5.4 and gpt-5.5 "},"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":"2607.15439","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-16T20:18:59Z","cross_cats_sorted":[],"title_canon_sha256":"65a4ca41a07fd49d0748c0cd1d69e916290b058f9adb9370a97c40eaea0f94ae","abstract_canon_sha256":"b730937362dd8a735b7976707c8a137a62444028dc582032f09fb3afb79f189b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-20T00:18:34.159843Z","signature_b64":"tiI/ilVmev5l/+6fdt7hKA223RUjJ0jAvRbNpZb6YhXpmQPSFMYWKs7F30reHdxylIEL8PsWcNtHMcaiUy1oCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"210c0dbb2e75ac82b4452103fd9ed1768d0a49e23c648efe7ede5af746a7f7c1","last_reissued_at":"2026-07-20T00:18:34.159055Z","signature_status":"signed_v1","first_computed_at":"2026-07-20T00:18:34.159055Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Coding Agents Need Executable World Models, Simplification, and Verification to Solve ARC-AGI-3?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Sergey Rodionov","submitted_at":"2026-07-16T20:18:59Z","abstract_excerpt":"Our previous ARC-AGI-3 agent bundled executable world modeling, scheduled simplification, and exact replay verification, leaving unclear which idea accounted for its performance. We address this attribution question with four nested Codex-based agents: a textual baseline; a flexible-interface executable world model without replay verification; the same executable model with scheduled simplification; and a fixed-interface verification treatment that retains simplification and requires exact reproduction of recorded observations. The main study evaluates all four agents with gpt-5.4 and gpt-5.5 "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.15439","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/2607.15439/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":"2607.15439","created_at":"2026-07-20T00:18:34.159447+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.15439v1","created_at":"2026-07-20T00:18:34.159447+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.15439","created_at":"2026-07-20T00:18:34.159447+00:00"},{"alias_kind":"pith_short_12","alias_value":"EEGA3OZOOWWI","created_at":"2026-07-20T00:18:34.159447+00:00"},{"alias_kind":"pith_short_16","alias_value":"EEGA3OZOOWWIFNCF","created_at":"2026-07-20T00:18:34.159447+00:00"},{"alias_kind":"pith_short_8","alias_value":"EEGA3OZO","created_at":"2026-07-20T00:18:34.159447+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/EEGA3OZOOWWIFNCFEEB73HWRO2","json":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2.json","graph_json":"https://pith.science/api/pith-number/EEGA3OZOOWWIFNCFEEB73HWRO2/graph.json","events_json":"https://pith.science/api/pith-number/EEGA3OZOOWWIFNCFEEB73HWRO2/events.json","paper":"https://pith.science/paper/EEGA3OZO"},"agent_actions":{"view_html":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2","download_json":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2.json","view_paper":"https://pith.science/paper/EEGA3OZO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.15439&json=true","fetch_graph":"https://pith.science/api/pith-number/EEGA3OZOOWWIFNCFEEB73HWRO2/graph.json","fetch_events":"https://pith.science/api/pith-number/EEGA3OZOOWWIFNCFEEB73HWRO2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2/action/storage_attestation","attest_author":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2/action/author_attestation","sign_citation":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2/action/citation_signature","submit_replication":"https://pith.science/pith/EEGA3OZOOWWIFNCFEEB73HWRO2/action/replication_record"}},"created_at":"2026-07-20T00:18:34.159447+00:00","updated_at":"2026-07-20T00:18:34.159447+00:00"}