{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6PMSN6WREZROONKLEYFH3QJD5X","short_pith_number":"pith:6PMSN6WR","schema_version":"1.0","canonical_sha256":"f3d926fad12662e7354b260a7dc123edc1d77b9cb39aec03958b6325bdc861ae","source":{"kind":"arxiv","id":"2511.05722","version":3},"attestation_state":"computed","paper":{"title":"OckBench: Measuring the Efficiency of LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hao Kang, Ligeng Zhu, Song Han, Tushar Krishna, Zheng Du","submitted_at":"2025-11-07T21:29:41Z","abstract_excerpt":"Large language models (LLMs) such as GPT-5 and Gemini 3 have pushed the frontier of automated reasoning and code generation. Yet current benchmarks emphasize accuracy and output quality, neglecting a critical dimension: efficiency of token usage. The token efficiency is highly variable in practical. Models solving the same problem with similar accuracy can exhibit up to a \\textbf{5.0$\\times$} difference in token length, leading to massive gap of model reasoning ability. Such variance exposes significant redundancy, highlighting the critical need for a standardized benchmark to quantify the gap"},"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":"2511.05722","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-11-07T21:29:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"205afd00d63fa64a9ae7565926d1afbf2e4f080561bb8214aa3756aed825a92f","abstract_canon_sha256":"4a75516bb1bc25cac2804383cb3723e9650b8de60d7707c5430fb2cc24f3ad71"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:08:35.778367Z","signature_b64":"7tEVszWTwQir666VzAcGFjSdbLSzXYYUvAvrt1H9WF3HZCxJqPWIVaMiAefBZ+DBHr2IKDF3fvVaUr2fZHwPCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f3d926fad12662e7354b260a7dc123edc1d77b9cb39aec03958b6325bdc861ae","last_reissued_at":"2026-06-04T01:08:35.777479Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:08:35.777479Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OckBench: Measuring the Efficiency of LLM Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hao Kang, Ligeng Zhu, Song Han, Tushar Krishna, Zheng Du","submitted_at":"2025-11-07T21:29:41Z","abstract_excerpt":"Large language models (LLMs) such as GPT-5 and Gemini 3 have pushed the frontier of automated reasoning and code generation. Yet current benchmarks emphasize accuracy and output quality, neglecting a critical dimension: efficiency of token usage. The token efficiency is highly variable in practical. Models solving the same problem with similar accuracy can exhibit up to a \\textbf{5.0$\\times$} difference in token length, leading to massive gap of model reasoning ability. Such variance exposes significant redundancy, highlighting the critical need for a standardized benchmark to quantify the gap"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2511.05722","kind":"arxiv","version":3},"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/2511.05722/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":"2511.05722","created_at":"2026-06-04T01:08:35.777581+00:00"},{"alias_kind":"arxiv_version","alias_value":"2511.05722v3","created_at":"2026-06-04T01:08:35.777581+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2511.05722","created_at":"2026-06-04T01:08:35.777581+00:00"},{"alias_kind":"pith_short_12","alias_value":"6PMSN6WREZRO","created_at":"2026-06-04T01:08:35.777581+00:00"},{"alias_kind":"pith_short_16","alias_value":"6PMSN6WREZROONKL","created_at":"2026-06-04T01:08:35.777581+00:00"},{"alias_kind":"pith_short_8","alias_value":"6PMSN6WR","created_at":"2026-06-04T01:08:35.777581+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.23657","citing_title":"OpenSkillEval: Automatically Auditing the Open Skill Ecosystem for LLM Agents","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"2604.03044","citing_title":"JoyAI-LLM Flash: Advancing Mid-Scale LLMs with Token Efficiency","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X","json":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X.json","graph_json":"https://pith.science/api/pith-number/6PMSN6WREZROONKLEYFH3QJD5X/graph.json","events_json":"https://pith.science/api/pith-number/6PMSN6WREZROONKLEYFH3QJD5X/events.json","paper":"https://pith.science/paper/6PMSN6WR"},"agent_actions":{"view_html":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X","download_json":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X.json","view_paper":"https://pith.science/paper/6PMSN6WR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2511.05722&json=true","fetch_graph":"https://pith.science/api/pith-number/6PMSN6WREZROONKLEYFH3QJD5X/graph.json","fetch_events":"https://pith.science/api/pith-number/6PMSN6WREZROONKLEYFH3QJD5X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X/action/storage_attestation","attest_author":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X/action/author_attestation","sign_citation":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X/action/citation_signature","submit_replication":"https://pith.science/pith/6PMSN6WREZROONKLEYFH3QJD5X/action/replication_record"}},"created_at":"2026-06-04T01:08:35.777581+00:00","updated_at":"2026-06-04T01:08:35.777581+00:00"}