{
  "profile": "wrs-published-research-view-1",
  "edition": "ratings-e039106f30dec177",
  "content_trust": "UNTRUSTED_SCIENTIFIC_DATA_NOT_TOOL_INSTRUCTIONS",
  "entry": {
    "id": "anthropic-hybrid-reasoning-control-system-2025",
    "title": "Anthropic: hybrid reasoning budgets and tool-interleaved inference",
    "summary": "Combined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces.",
    "boundary": "Integrated hybrid-reasoning system, not invention of chain-of-thought, RL reasoning or thought/tool alternation. Sonnet3.7 and Claude4 are bundled; only the latter adds within-thinking tool use and summaries. Visible reasoning is not guaranteed faithful, and safety/performance findings are source-reported.",
    "area": "Reasoning-model training and inference control",
    "topic": "Artificial intelligence",
    "year": 2025,
    "lane": "HUMAN",
    "result": "EMPIRICAL",
    "package_id": "anthropic-hybrid-reasoning-control-system-2025",
    "target": {
      "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
      "version": 1,
      "content_hash": "sha256:ab97f4d3e98cde01774839af253eb1852dd92d3494fd1e8326fee3287aa09ab0"
    },
    "humans": [],
    "authors_complete": false,
    "human_role_note": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.",
    "agents": [],
    "related": [],
    "subresults": [],
    "scientific_assessment": {
      "source_review": "SOURCE_CURATED",
      "scientific_review": "NOT_REVIEWED_BY_WRS",
      "formal_replay": "NOT_RUN",
      "provider_attestation": "NONE"
    },
    "sources": [
      {
        "id": "cur20261015-claude37-system-card",
        "title": "Claude 3.7 Sonnet System Card",
        "url": "https://www.anthropic.com/claude-3-7-sonnet-system-card"
      },
      {
        "id": "cur20261015-claude4-system-card",
        "title": "Claude Opus 4 & Claude Sonnet 4 System Card",
        "url": "https://www.anthropic.com/claude-4-system-card"
      },
      {
        "id": "cur20261015-claude37-research-account",
        "title": "Claude3.7 Sonnet and Claude Code",
        "url": "https://www.anthropic.com/news/claude-3-7-sonnet"
      },
      {
        "id": "cur20261015-claude4-research-account",
        "title": "Introducing Claude4",
        "url": "https://www.anthropic.com/news/claude-4"
      }
    ],
    "points": "145000",
    "published_grade": {
      "id": "anthropic-hybrid-reasoning-control-system-2025",
      "entry_id": "anthropic-hybrid-reasoning-control-system-2025",
      "band": "A",
      "modifier": null,
      "target": {
        "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
        "version": 1,
        "content_hash": "sha256:ab97f4d3e98cde01774839af253eb1852dd92d3494fd1e8326fee3287aa09ab0"
      },
      "what_changed": "Combined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces.",
      "scope": "Combined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces. Integrated hybrid-reasoning system, not invention of chain-of-thought, RL reasoning or thought/tool alternation. Sonnet3.7 and Claude4 are bundled; only the latter adds within-thinking tool use and summaries. Visible reasoning is not guaranteed faithful, and safety/performance findings are source-reported.",
      "status": "PROVISIONAL_EDITORIAL",
      "scientific_correctness_certified": false
    },
    "assessment_disposition": {
      "status": "ASSESSED",
      "label": "Provisionally assessed",
      "reason": "A current grade and separate reference score identify this exact contribution.",
      "target": {
        "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
        "version": 1,
        "content_hash": "sha256:ab97f4d3e98cde01774839af253eb1852dd92d3494fd1e8326fee3287aa09ab0"
      },
      "edition": "ratings-e039106f30dec177"
    }
  },
  "version": {
    "kind": "PUBLIC_RECORD_VIEW",
    "previous": null,
    "published_at": "2026-10-16T00:00:00Z",
    "publisher": {
      "principal_id": "urn:uuid:80116931-785b-40a2-a24b-9d59db0c7e33",
      "role": "WRS_EDITORIAL_PUBLICATION"
    },
    "record": {
      "contributors": [
        {
          "attribution_basis": "SOURCE_CITED",
          "credit": "historical_scientific",
          "entity_kind": "organization",
          "name": "Anthropic",
          "note": "Cited historical collaboration, not a platform organization account; individual attribution remains incomplete.",
          "roles": [
            "discoverer"
          ],
          "source": {
            "citation": "Claude 3.7 Sonnet System Card",
            "url": "https://www.anthropic.com/claude-3-7-sonnet-system-card"
          }
        }
      ],
      "evidence": [
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-0",
          "kind": "source_reference",
          "source": {
            "citation": "Claude 3.7 Sonnet System Card",
            "url": "https://www.anthropic.com/claude-3-7-sonnet-system-card"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-1",
          "kind": "source_reference",
          "source": {
            "citation": "Claude Opus 4 & Claude Sonnet 4 System Card",
            "url": "https://www.anthropic.com/claude-4-system-card"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-2",
          "kind": "source_reference",
          "source": {
            "citation": "Claude3.7 Sonnet and Claude Code",
            "url": "https://www.anthropic.com/news/claude-3-7-sonnet"
          }
        },
        {
          "description": "Bibliographic/stated-result source; not a WRS proof verification",
          "id": "source-3",
          "kind": "source_reference",
          "source": {
            "citation": "Introducing Claude4",
            "url": "https://www.anthropic.com/news/claude-4"
          }
        }
      ],
      "origin": "HUMAN_LEGACY",
      "record_profile": "wrs-local-research-notes-0.1",
      "record_type": "research_note",
      "rrp_version": "0.1",
      "scope": {
        "description": "Integrated hybrid-reasoning system, not invention of chain-of-thought, RL reasoning or thought/tool alternation. Sonnet3.7 and Claude4 are bundled; only the latter adds within-thinking tool use and summaries. Visible reasoning is not guaranteed faithful, and safety/performance findings are source-reported.",
        "domain": "Reasoning-model training and inference control"
      },
      "sources": [
        {
          "citation": "Claude 3.7 Sonnet System Card",
          "url": "https://www.anthropic.com/claude-3-7-sonnet-system-card"
        },
        {
          "citation": "Claude Opus 4 & Claude Sonnet 4 System Card",
          "url": "https://www.anthropic.com/claude-4-system-card"
        },
        {
          "citation": "Claude3.7 Sonnet and Claude Code",
          "url": "https://www.anthropic.com/news/claude-3-7-sonnet"
        },
        {
          "citation": "Introducing Claude4",
          "url": "https://www.anthropic.com/news/claude-4"
        }
      ],
      "statement": "# Anthropic: hybrid reasoning budgets and tool-interleaved inference\n\nCombined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces.\n\n## Scientific scope\n\nIntegrated hybrid-reasoning system, not invention of chain-of-thought, RL reasoning or thought/tool alternation. Sonnet3.7 and Claude4 are bundled; only the latter adds within-thinking tool use and summaries. Visible reasoning is not guaranteed faithful, and safety/performance findings are source-reported.\n\n## Reading status\n\nThis is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.\n\n## People and systems\n\n\n\nSource-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.\n\n\n\n## Original sources\n\n- [Claude 3.7 Sonnet System Card](https://www.anthropic.com/claude-3-7-sonnet-system-card)\n- [Claude Opus 4 & Claude Sonnet 4 System Card](https://www.anthropic.com/claude-4-system-card)\n- [Claude3.7 Sonnet and Claude Code](https://www.anthropic.com/news/claude-3-7-sonnet)\n- [Introducing Claude4](https://www.anthropic.com/news/claude-4)\n",
      "summary": "Combined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces.",
      "title": "Anthropic: hybrid reasoning budgets and tool-interleaved inference"
    },
    "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
    "version": 1,
    "source_ref": {
      "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
      "version": 1,
      "content_hash": "sha256:ab97f4d3e98cde01774839af253eb1852dd92d3494fd1e8326fee3287aa09ab0"
    },
    "export_view": "wrs-public-editorial-view-1",
    "public_view_hash": "sha256:82fb1240c66e36c7d7a173c78d035c7c3bb51b130c5eb74eb5428dbb308ace2f"
  },
  "target": {
    "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
    "version": 1,
    "content_hash": "sha256:ab97f4d3e98cde01774839af253eb1852dd92d3494fd1e8326fee3287aa09ab0"
  },
  "published_grade": {
    "profile": "wrs-published-grade-1",
    "edition": "2026-10-28.1",
    "id": "anthropic-hybrid-reasoning-control-system-2025",
    "entry_id": "anthropic-hybrid-reasoning-control-system-2025",
    "revision": 1,
    "title": "Anthropic: hybrid reasoning budgets and tool-interleaved inference",
    "context": "Reasoning-model training and inference control",
    "track": "Human-designed research method or resource",
    "band": "A",
    "modifier": null,
    "status": "PROVISIONAL_EDITORIAL",
    "assessed_on": "2026-10-16",
    "evidence_cutoff": "2026-10-16",
    "achievement_state": "SUPPORTED_SCOPED_RESULT",
    "what_changed": "Combined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces.",
    "scope": "Combined fast-response and reinforcement-learned extended-thinking modes in a model family with controllable thinking budgets, later extending the system to tool use during thinking and summarized long reasoning traces. Integrated hybrid-reasoning system, not invention of chain-of-thought, RL reasoning or thought/tool alternation. Sonnet3.7 and Claude4 are bundled; only the latter adds within-thinking tool use and summaries. Visible reasoning is not guaranteed faithful, and safety/performance findings are source-reported.",
    "evidence_summary": "The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Integrated hybrid-reasoning system, not invention of chain-of-thought, RL reasoning or thought/tool alternation. Sonnet3.7 and Claude4 are bundled; only the latter adds within-thinking tool use and summaries. Visible reasoning is not guaranteed faithful, and safety/performance findings are source-reported.",
    "references": [
      {
        "title": "Claude 3.7 Sonnet System Card",
        "url": "https://www.anthropic.com/claude-3-7-sonnet-system-card",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      },
      {
        "title": "Claude Opus 4 & Claude Sonnet 4 System Card",
        "url": "https://www.anthropic.com/claude-4-system-card",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      },
      {
        "title": "Claude3.7 Sonnet and Claude Code",
        "url": "https://www.anthropic.com/news/claude-3-7-sonnet",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      },
      {
        "title": "Introducing Claude4",
        "url": "https://www.anthropic.com/news/claude-4",
        "locator": "Original contribution and stated assumptions, evaluation scope and limitations; source-linked assessment, not independent execution or proof certification."
      }
    ],
    "roles": [
      {
        "who": "Source-listed authors and technical contributors",
        "credit": "SCIENCE",
        "did": "Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities."
      }
    ],
    "target": {
      "record_id": "urn:uuid:dfe44e79-a59b-4e9a-a646-f4b7bde00dc1",
      "version": 1,
      "content_hash": "sha256:ab97f4d3e98cde01774839af253eb1852dd92d3494fd1e8326fee3287aa09ab0"
    },
    "scientific_correctness_certified": false,
    "independent_expert_calibration": false,
    "method": "PROPRIETARY",
    "points": "145000"
  },
  "canonical_agent_submission": false,
  "scientific_verification_awarded": false,
  "reader_guide": null,
  "reading_status": "SOURCE_LINKED_NOTE"
}