{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BAJ7IIXD4C6K726PNXOUHD3EZT","short_pith_number":"pith:BAJ7IIXD","schema_version":"1.0","canonical_sha256":"0813f422e3e0bcafebcf6ddd438f64ccd786fdeecaa46bf4ca50fca7784135b8","source":{"kind":"arxiv","id":"2410.12049","version":4},"attestation_state":"computed","paper":{"title":"Sabi\\'a-3 Technical Report","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Giovana Kerche Bon\\'as, Hugo Abonizio, Ramon Pires, Rodrigo Nogueira, Roseval Malaquias Junior, Thales Sales Almeida, Thiago Laitz","submitted_at":"2024-10-15T20:37:34Z","abstract_excerpt":"This report presents Sabi\\'a-3, our new flagship language model, and Sabiazinho-3, a more cost-effective sibling. The models were trained on a large brazilian-centric corpus. Evaluations across diverse professional and academic benchmarks show a strong performance on Portuguese and Brazil-related tasks. Sabi\\'a-3 shows large improvements in comparison to our previous best of model, Sabia-2 Medium, especially in reasoning-intensive tasks. Notably, Sabi\\'a-3's average performance matches frontier LLMs, while it is offered at a three to four times lower cost per token, reinforcing the benefits of"},"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":"2410.12049","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-15T20:37:34Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2e3556024be2b481448d539e3b80afe620b23c3a1acdf85af3a57e9329added7","abstract_canon_sha256":"ef9c0071dc1ba47dd9988806080e8ed903a5d3613166b3a8cfc9463d8e0d428d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:42:38.273614Z","signature_b64":"Zf+U/qgbbrZKHBgPEKRV+g4yPwFED3XsZUtuDwK3xT+m4f3h5An0Rb5j7O7LxXJPurI6uabFp8rcW8JcD7sYCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0813f422e3e0bcafebcf6ddd438f64ccd786fdeecaa46bf4ca50fca7784135b8","last_reissued_at":"2026-07-05T10:42:38.273085Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:42:38.273085Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sabi\\'a-3 Technical Report","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Giovana Kerche Bon\\'as, Hugo Abonizio, Ramon Pires, Rodrigo Nogueira, Roseval Malaquias Junior, Thales Sales Almeida, Thiago Laitz","submitted_at":"2024-10-15T20:37:34Z","abstract_excerpt":"This report presents Sabi\\'a-3, our new flagship language model, and Sabiazinho-3, a more cost-effective sibling. The models were trained on a large brazilian-centric corpus. Evaluations across diverse professional and academic benchmarks show a strong performance on Portuguese and Brazil-related tasks. Sabi\\'a-3 shows large improvements in comparison to our previous best of model, Sabia-2 Medium, especially in reasoning-intensive tasks. Notably, Sabi\\'a-3's average performance matches frontier LLMs, while it is offered at a three to four times lower cost per token, reinforcing the benefits of"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.12049","kind":"arxiv","version":4},"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/2410.12049/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":"2410.12049","created_at":"2026-07-05T10:42:38.273148+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.12049v4","created_at":"2026-07-05T10:42:38.273148+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.12049","created_at":"2026-07-05T10:42:38.273148+00:00"},{"alias_kind":"pith_short_12","alias_value":"BAJ7IIXD4C6K","created_at":"2026-07-05T10:42:38.273148+00:00"},{"alias_kind":"pith_short_16","alias_value":"BAJ7IIXD4C6K726P","created_at":"2026-07-05T10:42:38.273148+00:00"},{"alias_kind":"pith_short_8","alias_value":"BAJ7IIXD","created_at":"2026-07-05T10:42:38.273148+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2512.15791","citing_title":"Evaluation of AI Ethics Tools in Language Models: A Developers' Perspective Case Study","ref_index":101,"is_internal_anchor":false},{"citing_arxiv_id":"2601.18622","citing_title":"Brazilian Social Media Anti-vaccine Information Disorder Dataset -- Telegram (2020-2025)","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT","json":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT.json","graph_json":"https://pith.science/api/pith-number/BAJ7IIXD4C6K726PNXOUHD3EZT/graph.json","events_json":"https://pith.science/api/pith-number/BAJ7IIXD4C6K726PNXOUHD3EZT/events.json","paper":"https://pith.science/paper/BAJ7IIXD"},"agent_actions":{"view_html":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT","download_json":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT.json","view_paper":"https://pith.science/paper/BAJ7IIXD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.12049&json=true","fetch_graph":"https://pith.science/api/pith-number/BAJ7IIXD4C6K726PNXOUHD3EZT/graph.json","fetch_events":"https://pith.science/api/pith-number/BAJ7IIXD4C6K726PNXOUHD3EZT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT/action/storage_attestation","attest_author":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT/action/author_attestation","sign_citation":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT/action/citation_signature","submit_replication":"https://pith.science/pith/BAJ7IIXD4C6K726PNXOUHD3EZT/action/replication_record"}},"created_at":"2026-07-05T10:42:38.273148+00:00","updated_at":"2026-07-05T10:42:38.273148+00:00"}