{"paper":{"title":"SciResearcher: Scaling Deep Research Agents for Frontier Scientific Reasoning","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"Automated synthesis of conceptual and computational tasks trains an 8B model to set new records on frontier biology and chemistry reasoning benchmarks.","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Kelvin Kiu Wai Tam, Newt Nguyen Kim Hue Nam, Rui Wang, Tianqing Fang, Tianshi Zheng, Wei Fan, Xiyun Li, Yangqiu Song","submitted_at":"2026-05-02T15:26:45Z","abstract_excerpt":"Frontier scientific reasoning is rapidly emerging as a key foundation for advancing AI agents in automated scientific discovery. Deep research agents offer a promising approach to this challenge. These models develop robust problem-solving capabilities through post-training on information-seeking tasks, which are typically curated via knowledge graph construction or iterative web browsing. However, these strategies face inherent limitations in frontier science, where domain-specific knowledge is scattered across sparse and heterogeneous academic sources, and problem solving requires sophistica"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SciResearcher-8B achieves 19.46% on the HLE-Bio/Chem-Gold benchmark, establishing a new state of the art at its parameter scale and surpassing several larger proprietary agents. It further achieves 13-15% absolute gains on SuperGPQA-Hard-Biology and TRQA-Literature benchmarks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That tasks synthesized by the agentic framework accurately reflect the computational and reasoning demands of actual frontier scientific problems rather than simplified or proxy versions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SciResearcher automates creation of diverse scientific reasoning tasks from academic evidence to train an 8B model that sets new SOTA at 19.46% on HLE-Bio/Chem-Gold and gains 13-15% on SuperGPQA-Hard-Biology and TRQA-Literature.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Automated synthesis of conceptual and computational tasks trains an 8B model to set new records on frontier biology and chemistry reasoning benchmarks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"9a738753c9f2398ed340feb06f96acd02965d24f57e23cab512eaa68a7072ac8"},"source":{"id":"2605.01489","kind":"arxiv","version":2},"verdict":{"id":"b0632074-89ac-46f4-82d8-5c9660c72ff8","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T14:14:38.432631Z","strongest_claim":"SciResearcher-8B achieves 19.46% on the HLE-Bio/Chem-Gold benchmark, establishing a new state of the art at its parameter scale and surpassing several larger proprietary agents. It further achieves 13-15% absolute gains on SuperGPQA-Hard-Biology and TRQA-Literature benchmarks.","one_line_summary":"SciResearcher automates creation of diverse scientific reasoning tasks from academic evidence to train an 8B model that sets new SOTA at 19.46% on HLE-Bio/Chem-Gold and gains 13-15% on SuperGPQA-Hard-Biology and TRQA-Literature.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That tasks synthesized by the agentic framework accurately reflect the computational and reasoning demands of actual frontier scientific problems rather than simplified or proxy versions.","pith_extraction_headline":"Automated synthesis of conceptual and computational tasks trains an 8B model to set new records on frontier biology and chemistry reasoning benchmarks."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.01489/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T17:41:04.058433Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T17:13:33.807168Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"68abf098f1144e9428abf2bf1bf488b5638eb76e9dda3d6b9a6fa19ddb4d447e"},"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"}