{
  "$schema": "https://themachinepress.com/schemas/story-v1.schema.json",
  "schema_version": "1.0.0",
  "document_type": "machine_press_story",
  "story": {
    "story_id": "mp-2026-11-12-019",
    "source_story_id": "tmp-brief-bioxarena-biomedical-agents",
    "edition_id": "mp-2026-11-12-morning-0125",
    "edition_url": "https://themachinepress.com/edition/2026-11-12",
    "position": 21,
    "story_type": "ticker",
    "section": "benchmarks-evals",
    "editorial_classification": "editorial",
    "headline": "Biomedical Agents Meet Rebuilt Research Tasks",
    "slug": "biomedical-agents-meet-rebuilt-research-tasks",
    "dek": "BioXArena-Pro holds out biological units and finds that scaffolding matters nearly as much as the model.",
    "summary": "BioXArena-Pro holds out biological units and finds that scaffolding matters nearly as much as the model.",
    "body_text": "BioXArena-Pro rebuilds 64 modeling tasks across eight biomedical domains, usually holding out whole species or cohorts. Across 22 agent configurations the best score was 55.1 of 100, and removing pretrained biological models cost nine points on tasks where the leading agent used them.",
    "why_it_matters": "BioXArena-Pro holds out biological units and finds that scaffolding matters nearly as much as the model.",
    "limitations": [],
    "importance": 7,
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-12-019/biomedical-agents-meet-rebuilt-research-tasks",
    "json_url": "https://themachinepress.com/story/mp-2026-11-12-019.json",
    "first_published_at": "2026-11-12T09:00:00.000-05:00",
    "modified_at": "2026-11-12T09:00:00.000-05:00",
    "content_status": "new",
    "is_carryover": false,
    "carryover_reason": null,
    "key_claims": [
      {
        "claim_id": "claim-mp-2026-11-12-019-001",
        "text": "BioXArena-Pro holds out biological units and finds that scaffolding matters nearly as much as the model.",
        "source_ids": [
          "source-2026-11-12-021"
        ],
        "qualification": null
      }
    ],
    "source_ids": [
      "source-2026-11-12-021"
    ],
    "tags": [
      "biomedical",
      "agents",
      "benchmark",
      "modeling"
    ],
    "image_url": null,
    "corrections": []
  },
  "sources": [
    {
      "source_id": "source-2026-11-12-021",
      "title": "arXiv:2611.13994v1",
      "publisher": "arXiv / Cornell University",
      "url": "https://arxiv.org/abs/2611.13994v1",
      "canonical_url": "https://arxiv.org/abs/2611.13994v1",
      "source_type": "primary_research",
      "is_primary_source": true,
      "published_at": "2026-11-10T07:39:35.000-05:00",
      "accessed_at": "2026-11-12T08:24:40.465-05:00",
      "supports_claim_ids": [
        "claim-mp-2026-11-12-019-001"
      ]
    }
  ],
  "corrections": [],
  "publisher": {
    "name": "The Machine Press",
    "url": "https://themachinepress.com",
    "description": "A daily newspaper for the age of artificial intelligence."
  },
  "cite_this_report": {
    "title": "Biomedical Agents Meet Rebuilt Research Tasks",
    "publisher": "The Machine Press",
    "published_at": "2026-11-12T09:00:00.000-05:00",
    "canonical_url": "https://themachinepress.com/story/mp-2026-11-12-019/biomedical-agents-meet-rebuilt-research-tasks"
  }
}