{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JOW3CFTNXIFMQCXMBTYEQ5JRFO","short_pith_number":"pith:JOW3CFTN","schema_version":"1.0","canonical_sha256":"4badb1166dba0ac80aec0cf04875312b8de5154bedcd1e1a8598fea68537a98d","source":{"kind":"arxiv","id":"2606.00579","version":1},"attestation_state":"computed","paper":{"title":"Sandboxed Coding Agents are Competitive Omni-modal Task Solvers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Dianqi Li, Dongping Chen, Qingyuan Shi, Tianyi Zhou, Xuanao Huang, Zhihan Hu","submitted_at":"2026-05-30T07:04:27Z","abstract_excerpt":"As multimodal LLMs increasingly target video and audio, it is often assumed that such tasks require native omnimodal models. We show that this is not always the case: coding agents with only text+image access and a sandboxed tool-use interface can match, and in several settings outperform, SOTA native omnimodal models and predefined multimodal agent scaffolds across multiple audio-video benchmarks. Our trajectory analysis suggests that their strength comes from writing code and orchestrating tools to extract relevant evidence from transcripts, frames, and other modality signals, thereby conver"},"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":"2606.00579","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-30T07:04:27Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"5be662cec7385b4569a93a7daadc8fb49f7b43f52a0142b48f2bc489ddc74cc8","abstract_canon_sha256":"3788a4cd6135db6dfbb3845ffc2caa2a9687c23eeca29903508fe795cde1bc32"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:58.814131Z","signature_b64":"M/ahHFsIUX6JZq4gQLRnjOBTjZXDj9ipB6FZuPsd70tvTXioQ+T+Czdnyebp+OODoTF57ONisaKDCQmQXRZjDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4badb1166dba0ac80aec0cf04875312b8de5154bedcd1e1a8598fea68537a98d","last_reissued_at":"2026-06-02T01:03:58.813730Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:58.813730Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sandboxed Coding Agents are Competitive Omni-modal Task Solvers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Dianqi Li, Dongping Chen, Qingyuan Shi, Tianyi Zhou, Xuanao Huang, Zhihan Hu","submitted_at":"2026-05-30T07:04:27Z","abstract_excerpt":"As multimodal LLMs increasingly target video and audio, it is often assumed that such tasks require native omnimodal models. We show that this is not always the case: coding agents with only text+image access and a sandboxed tool-use interface can match, and in several settings outperform, SOTA native omnimodal models and predefined multimodal agent scaffolds across multiple audio-video benchmarks. Our trajectory analysis suggests that their strength comes from writing code and orchestrating tools to extract relevant evidence from transcripts, frames, and other modality signals, thereby conver"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00579","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/2606.00579/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":"2606.00579","created_at":"2026-06-02T01:03:58.813782+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00579v1","created_at":"2026-06-02T01:03:58.813782+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00579","created_at":"2026-06-02T01:03:58.813782+00:00"},{"alias_kind":"pith_short_12","alias_value":"JOW3CFTNXIFM","created_at":"2026-06-02T01:03:58.813782+00:00"},{"alias_kind":"pith_short_16","alias_value":"JOW3CFTNXIFMQCXM","created_at":"2026-06-02T01:03:58.813782+00:00"},{"alias_kind":"pith_short_8","alias_value":"JOW3CFTN","created_at":"2026-06-02T01:03:58.813782+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/JOW3CFTNXIFMQCXMBTYEQ5JRFO","json":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO.json","graph_json":"https://pith.science/api/pith-number/JOW3CFTNXIFMQCXMBTYEQ5JRFO/graph.json","events_json":"https://pith.science/api/pith-number/JOW3CFTNXIFMQCXMBTYEQ5JRFO/events.json","paper":"https://pith.science/paper/JOW3CFTN"},"agent_actions":{"view_html":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO","download_json":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO.json","view_paper":"https://pith.science/paper/JOW3CFTN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00579&json=true","fetch_graph":"https://pith.science/api/pith-number/JOW3CFTNXIFMQCXMBTYEQ5JRFO/graph.json","fetch_events":"https://pith.science/api/pith-number/JOW3CFTNXIFMQCXMBTYEQ5JRFO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO/action/storage_attestation","attest_author":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO/action/author_attestation","sign_citation":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO/action/citation_signature","submit_replication":"https://pith.science/pith/JOW3CFTNXIFMQCXMBTYEQ5JRFO/action/replication_record"}},"created_at":"2026-06-02T01:03:58.813782+00:00","updated_at":"2026-06-02T01:03:58.813782+00:00"}