{
  "abstract": "Domain-specialized language models are widely used for scientific question answering, but stronger general-purpose systems raise a sharper question: when does domain-specific fine-tuning remain valuable for open-ended scientific reasoning? We study this in astronomy with a curated QA benchmark from publicly available 2017--2026 Olympiad-style materials. The free-response subset contains 300 questions, including 204 text-only and 96 image-linked examples. We compare open-weight and API-served general-purpose, multimodal, and astronomy-specialized models using judge-based correctness and complementary reference metrics. Strong general-purpose models establish the highest correctness baseline in this testbed, while analyses of metric agreement, judge sensitivity, benchmark composition, and modality reveal variation not captured by a single leaderboard. These results motivate treating domain specialization as a task- and deployment-dependent property and highlight the role of domain-specific evaluation in determining which models, capabilities, and evaluation criteria are appropriate for scientific workflows.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.17644",
  "author_total": 8,
  "authors": [
    {
      "name": "Vanessa Lama"
    },
    {
      "name": "Sanjay Das"
    },
    {
      "name": "Emily Herron"
    },
    {
      "name": "Yuan-Sen Ting"
    },
    {
      "name": "Tijmen de Haan"
    },
    {
      "name": "Junqi Yin"
    },
    {
      "name": "Tirthankar Ghosal"
    },
    {
      "name": "Feiyi Wang"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/4cf2ntbm5h6hz.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.17644",
  "fulltext_url": null,
  "id": "4cf2ntbm5h6hz",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.17644",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "astro-ph.IM"
    ],
    [
      "arxiv",
      "cs.AI"
    ],
    [
      "arxiv",
      "cs.CY"
    ],
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "openalex",
      "1702|Topic Modeling"
    ]
  ],
  "openalex": "W7213513412",
  "pdf_url": "https://arxiv.org/pdf/2609.17644",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-15",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/4cf2ntbm5h6hz.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:22:01Z",
      "native_id": "2609.17644",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.17644"
    }
  ],
  "subjects": [
    {
      "field": "astronomy",
      "origin": "arxiv:astro-ph.IM",
      "subfield": "instrumentation-methods"
    }
  ],
  "title": "Rethinking Domain Specialization for Open-Ended Scientific Reasoning in Astronomy Language Models",
  "title_text": "Rethinking Domain Specialization for Open-Ended Scientific Reasoning in Astronomy Language Models",
  "type": "preprint",
  "updated_at": "2026-10-30T12:22:01Z",
  "url": "https://research.draconitic.com/paper/4cf2ntbm5h6hz",
  "venue": "arXiv",
  "versions": []
}