{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NTALJIUWD7FPIORBQZ3UDBJXUF","short_pith_number":"pith:NTALJIUW","schema_version":"1.0","canonical_sha256":"6cc0b4a2961fcaf43a218677418537a145f216e526e9e3ebe6325f904166d6cc","source":{"kind":"arxiv","id":"2605.14398","version":1},"attestation_state":"computed","paper":{"title":"Coding Agent Is Good As World Simulator","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A multi-agent framework generates and refines executable physics simulation code from prompts to create world models that enforce physical constraints, claiming superior accuracy and fidelity over video-based alternatives.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bocheng Zou, Dan Negrut, Hongyu Wang, Jingquan Wang, Radu Serban","submitted_at":"2026-05-14T05:33:41Z","abstract_excerpt":"World models have emerged as a powerful paradigm for building interactive simulation environments, with recent video-based approaches demonstrating impressive progress in generating visually plausible dynamics. However, because these models typically infer dynamics from video and represent them in latent states, they do not explicitly enforce physical constraints. As a result, the generated video rollouts are not physically plausible, exhibiting unstable contacts, distorted shapes, or inconsistent motion. In this paper, we present an agentic framework constructing physics-based world models th"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.14398","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-14T05:33:41Z","cross_cats_sorted":[],"title_canon_sha256":"313c79e996bdead27ead449ad311d39492bf69b6e714935fc1853a9d93162393","abstract_canon_sha256":"6a42d9f513213e04ba7c2e16b617ab18841d3da69bddd190a71b1b75f04cc54d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:07.521606Z","signature_b64":"nGki6aH30OgthhUO2zn08LDg19kjiedn1y8YA2fqkuoLm8YNXxVyLFten1NzpDOaAjRsDRgSSa2i/UMOIzZBCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6cc0b4a2961fcaf43a218677418537a145f216e526e9e3ebe6325f904166d6cc","last_reissued_at":"2026-05-17T23:39:07.520747Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:07.520747Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Coding Agent Is Good As World Simulator","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A multi-agent framework generates and refines executable physics simulation code from prompts to create world models that enforce physical constraints, claiming superior accuracy and fidelity over video-based alternatives.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bocheng Zou, Dan Negrut, Hongyu Wang, Jingquan Wang, Radu Serban","submitted_at":"2026-05-14T05:33:41Z","abstract_excerpt":"World models have emerged as a powerful paradigm for building interactive simulation environments, with recent video-based approaches demonstrating impressive progress in generating visually plausible dynamics. However, because these models typically infer dynamics from video and represent them in latent states, they do not explicitly enforce physical constraints. As a result, the generated video rollouts are not physically plausible, exhibiting unstable contacts, distorted shapes, or inconsistent motion. In this paper, we present an agentic framework constructing physics-based world models th"},"claims":{"count":3,"items":[{"kind":"strongest_claim","text":"Experimental results show that our framework outperforms advanced video-based models in physical accuracy, instruction fidelity and visual quality, which could be applied to various scenarios including driving simulation and embodied robot tasks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that the visual review and physics analysis agents can reliably detect and guide corrections for physical inconsistencies in generated code without ground-truth physics data or human intervention, allowing the iterative process to converge to valid simulations.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A multi-agent framework generates and refines executable physics simulation code from prompts to create world models that enforce physical constraints, claiming superior accuracy and fidelity over video-based alternatives.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"}],"snapshot_sha256":"35c13d353887cf1d5167bfc57d9da69c6df3ecc896799058acccddf50331b8d1"},"source":{"id":"2605.14398","kind":"arxiv","version":1},"verdict":{"id":"f7745108-3db8-478b-9a11-c1c4885ed48e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T02:15:46.574907Z","strongest_claim":"Experimental results show that our framework outperforms advanced video-based models in physical accuracy, instruction fidelity and visual quality, which could be applied to various scenarios including driving simulation and embodied robot tasks.","one_line_summary":"A multi-agent framework generates and refines executable physics simulation code from prompts to create world models that enforce physical constraints, claiming superior accuracy and fidelity over video-based alternatives.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that the visual review and physics analysis agents can reliably detect and guide corrections for physical inconsistencies in generated code without ground-truth physics data or human intervention, allowing the iterative process to converge to valid simulations.","pith_extraction_headline":""},"references":{"count":59,"sample":[{"doi":"","year":2018,"title":"Recurrent world models facilitate policy evolution","work_id":"d0afab3b-377f-4f25-b67f-7dc40edfab95","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"Learning latent dynamics for planning from pixels","work_id":"3ae77a86-0f76-42cd-9d88-1177a2bb1390","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"Dream to control: Learning behaviors by latent imagination","work_id":"0efe4973-b7ef-4bb2-9e89-b2b8436c3d2a","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Genie: Generative interactive environments","work_id":"6c633c28-756b-4f8a-b31e-d5ac37197f04","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"GAIA-1: A Generative World Model for Autonomous Driving","work_id":"313484e6-a442-4522-8e19-d07e502844a8","ref_index":5,"cited_arxiv_id":"2309.17080","is_internal_anchor":true}],"resolved_work":59,"snapshot_sha256":"4015c31735e7832463a2098aeee38157cee105655f85079f51878c01a1850240","internal_anchors":9},"formal_canon":{"evidence_count":2,"snapshot_sha256":"93d2c94f301d44d3cfb750aa814b0d9e148c4f2a27174c82880060c856a46a4f"},"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":"2605.14398","created_at":"2026-05-17T23:39:07.520885+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.14398v1","created_at":"2026-05-17T23:39:07.520885+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14398","created_at":"2026-05-17T23:39:07.520885+00:00"},{"alias_kind":"pith_short_12","alias_value":"NTALJIUWD7FP","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"NTALJIUWD7FPIORB","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"NTALJIUW","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF","json":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF.json","graph_json":"https://pith.science/api/pith-number/NTALJIUWD7FPIORBQZ3UDBJXUF/graph.json","events_json":"https://pith.science/api/pith-number/NTALJIUWD7FPIORBQZ3UDBJXUF/events.json","paper":"https://pith.science/paper/NTALJIUW"},"agent_actions":{"view_html":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF","download_json":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF.json","view_paper":"https://pith.science/paper/NTALJIUW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.14398&json=true","fetch_graph":"https://pith.science/api/pith-number/NTALJIUWD7FPIORBQZ3UDBJXUF/graph.json","fetch_events":"https://pith.science/api/pith-number/NTALJIUWD7FPIORBQZ3UDBJXUF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF/action/storage_attestation","attest_author":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF/action/author_attestation","sign_citation":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF/action/citation_signature","submit_replication":"https://pith.science/pith/NTALJIUWD7FPIORBQZ3UDBJXUF/action/replication_record"}},"created_at":"2026-05-17T23:39:07.520885+00:00","updated_at":"2026-05-17T23:39:07.520885+00:00"}