{
  "schema": "schemaxis.website-measured/1",
  "study": {
    "title": "Smaller models, trusted recipes and workflow authoring",
    "completed": "2026-09-29",
    "completedDisplay": "29 September 2026",
    "task": "ParaView workflow authoring and prepared-recipe operation on small, purpose-built serial ParaView test fixtures",
    "checksPerAuthoringConversation": 110,
    "initialChecks": 40,
    "revisionChecks": 70,
    "claudeReplicates": 3,
    "openaiReplicates": 1,
    "preparedTurns": 88,
    "preparedChecks": 4840,
    "preparedChecksDisplay": "4,840",
    "pricingRetrieved": "2026-09-29",
    "pricingRetrievedDisplay": "29 September 2026"
  },
  "authoring": [
    {
      "model": "Haiku 4.5",
      "modelKey": "haiku",
      "provider": "Anthropic",
      "tier": "smaller",
      "method": "native",
      "initial": "fail",
      "revision": "fail",
      "available": true,
      "checksPassed": 22,
      "checks": 110,
      "completePass": false,
      "seconds": 629.1765340708662,
      "costUSD": 1.0234461999999998,
      "rounds": {
        "initial": {
          "passed": 8,
          "total": 40,
          "imagesPassed": 0,
          "imagesTotal": 32,
          "imagesRejected": 32,
          "gradePassed": false
        },
        "revision": {
          "passed": 14,
          "total": 70,
          "imagesPassed": 0,
          "imagesTotal": 56,
          "imagesRejected": 56,
          "gradePassed": false
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "629.2",
        "secondsWhole": "629",
        "cost": "1.0234",
        "costShort": "1.02"
      }
    },
    {
      "model": "Haiku 4.5",
      "modelKey": "haiku",
      "provider": "Anthropic",
      "tier": "smaller",
      "method": "schemaxis",
      "initial": "pass",
      "revision": "pass",
      "available": true,
      "checksPassed": 110,
      "checks": 110,
      "completePass": true,
      "seconds": 827.5577170070028,
      "costUSD": 0.9423158000000003,
      "rounds": {
        "initial": {
          "passed": 40,
          "total": 40,
          "imagesPassed": 32,
          "imagesTotal": 32,
          "imagesRejected": 0,
          "gradePassed": true
        },
        "revision": {
          "passed": 70,
          "total": 70,
          "imagesPassed": 56,
          "imagesTotal": 56,
          "imagesRejected": 0,
          "gradePassed": true
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "827.6",
        "secondsWhole": "828",
        "cost": "0.9423",
        "costShort": "0.94"
      }
    },
    {
      "model": "Sonnet 4.6",
      "modelKey": "sonnet",
      "provider": "Anthropic",
      "tier": "smaller",
      "method": "native",
      "initial": "fail",
      "revision": "fail",
      "available": true,
      "checksPassed": 22,
      "checks": 110,
      "completePass": false,
      "seconds": 353.34096089494415,
      "costUSD": 0.8122500000000004,
      "rounds": {
        "initial": {
          "passed": 8,
          "total": 40,
          "imagesPassed": 0,
          "imagesTotal": 32,
          "imagesRejected": 32,
          "gradePassed": false
        },
        "revision": {
          "passed": 14,
          "total": 70,
          "imagesPassed": 0,
          "imagesTotal": 56,
          "imagesRejected": 56,
          "gradePassed": false
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "353.3",
        "secondsWhole": "353",
        "cost": "0.8123",
        "costShort": "0.81"
      }
    },
    {
      "model": "Sonnet 4.6",
      "modelKey": "sonnet",
      "provider": "Anthropic",
      "tier": "smaller",
      "method": "schemaxis",
      "initial": "pass",
      "revision": "pass",
      "available": true,
      "checksPassed": 110,
      "checks": 110,
      "completePass": true,
      "seconds": 757.1777740659891,
      "costUSD": 2.4922974,
      "rounds": {
        "initial": {
          "passed": 40,
          "total": 40,
          "imagesPassed": 32,
          "imagesTotal": 32,
          "imagesRejected": 0,
          "gradePassed": true
        },
        "revision": {
          "passed": 70,
          "total": 70,
          "imagesPassed": 56,
          "imagesTotal": 56,
          "imagesRejected": 0,
          "gradePassed": true
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "757.2",
        "secondsWhole": "757",
        "cost": "2.4923",
        "costShort": "2.49"
      }
    },
    {
      "model": "Opus 5.5",
      "modelKey": "opus",
      "provider": "Anthropic",
      "tier": "frontier",
      "method": "native",
      "initial": "unavailable",
      "revision": "not run",
      "available": false,
      "note": "Native control unavailable: the provider substituted another model, so the attempt was excluded.",
      "excludedReport": "ai-paraview-authoring-claude-01"
    },
    {
      "model": "Opus 5.5",
      "modelKey": "opus",
      "provider": "Anthropic",
      "tier": "frontier",
      "method": "schemaxis",
      "initial": "pass",
      "revision": "pass",
      "available": true,
      "checksPassed": 110,
      "checks": 110,
      "completePass": true,
      "seconds": 227.78997608402278,
      "costUSD": 1.1866856,
      "rounds": {
        "initial": {
          "passed": 40,
          "total": 40,
          "imagesPassed": 32,
          "imagesTotal": 32,
          "imagesRejected": 0,
          "gradePassed": true
        },
        "revision": {
          "passed": 70,
          "total": 70,
          "imagesPassed": 56,
          "imagesTotal": 56,
          "imagesRejected": 0,
          "gradePassed": true
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "227.8",
        "secondsWhole": "228",
        "cost": "1.1867",
        "costShort": "1.19"
      }
    },
    {
      "model": "GPT-5.6 Luna",
      "modelKey": "luna",
      "provider": "OpenAI",
      "tier": "smaller",
      "method": "native",
      "initial": "fail",
      "revision": "fail",
      "available": true,
      "checksPassed": 68,
      "checks": 110,
      "completePass": false,
      "seconds": 268.104292280972,
      "costUSD": 0.04706292000000001,
      "rounds": {
        "initial": {
          "passed": 26,
          "total": 40,
          "imagesPassed": 18,
          "imagesTotal": 32,
          "imagesRejected": 14,
          "gradePassed": false
        },
        "revision": {
          "passed": 42,
          "total": 70,
          "imagesPassed": 28,
          "imagesTotal": 56,
          "imagesRejected": 28,
          "gradePassed": false
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "268.1",
        "secondsWhole": "268",
        "cost": "0.0471",
        "costShort": "0.05"
      }
    },
    {
      "model": "GPT-5.6 Luna",
      "modelKey": "luna",
      "provider": "OpenAI",
      "tier": "smaller",
      "method": "schemaxis",
      "initial": "pass",
      "revision": "pass",
      "available": true,
      "checksPassed": 110,
      "checks": 110,
      "completePass": true,
      "seconds": 284.5720699230442,
      "costUSD": 0.08146432,
      "rounds": {
        "initial": {
          "passed": 40,
          "total": 40,
          "imagesPassed": 32,
          "imagesTotal": 32,
          "imagesRejected": 0,
          "gradePassed": true
        },
        "revision": {
          "passed": 70,
          "total": 70,
          "imagesPassed": 56,
          "imagesTotal": 56,
          "imagesRejected": 0,
          "gradePassed": true
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "284.6",
        "secondsWhole": "285",
        "cost": "0.0815",
        "costShort": "0.08"
      }
    },
    {
      "model": "GPT-6 Astra",
      "modelKey": "astra",
      "provider": "OpenAI",
      "tier": "frontier",
      "method": "native",
      "initial": "pass",
      "revision": "pass",
      "available": true,
      "checksPassed": 110,
      "checks": 110,
      "completePass": true,
      "seconds": 177.21313359902706,
      "costUSD": 0.882752,
      "rounds": {
        "initial": {
          "passed": 40,
          "total": 40,
          "imagesPassed": 32,
          "imagesTotal": 32,
          "imagesRejected": 0,
          "gradePassed": true
        },
        "revision": {
          "passed": 70,
          "total": 70,
          "imagesPassed": 56,
          "imagesTotal": 56,
          "imagesRejected": 0,
          "gradePassed": true
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "177.2",
        "secondsWhole": "177",
        "cost": "0.8828",
        "costShort": "0.88"
      }
    },
    {
      "model": "GPT-6 Astra",
      "modelKey": "astra",
      "provider": "OpenAI",
      "tier": "frontier",
      "method": "schemaxis",
      "initial": "pass",
      "revision": "pass",
      "available": true,
      "checksPassed": 110,
      "checks": 110,
      "completePass": true,
      "seconds": 305.01253737090155,
      "costUSD": 2.485264,
      "rounds": {
        "initial": {
          "passed": 40,
          "total": 40,
          "imagesPassed": 32,
          "imagesTotal": 32,
          "imagesRejected": 0,
          "gradePassed": true
        },
        "revision": {
          "passed": 70,
          "total": 70,
          "imagesPassed": 56,
          "imagesTotal": 56,
          "imagesRejected": 0,
          "gradePassed": true
        }
      },
      "conversations": 1,
      "display": {
        "seconds": "305.0",
        "secondsWhole": "305",
        "cost": "2.4853",
        "costShort": "2.49"
      }
    }
  ],
  "claims": {
    "smallerModels": {
      "models": [
        {
          "model": "Haiku 4.5",
          "native": "22/110",
          "schemaxis": "110/110"
        },
        {
          "model": "Sonnet 4.6",
          "native": "22/110",
          "schemaxis": "110/110"
        },
        {
          "model": "GPT-5.6 Luna",
          "native": "68/110",
          "schemaxis": "110/110"
        }
      ],
      "checks": 110,
      "conversationsPerCell": 1
    },
    "silentFailure": {
      "model": "Haiku 4.5",
      "imagesRejected": 32,
      "imagesTotal": 32,
      "geometryFieldPassed": 8,
      "initialPassed": 8,
      "initialTotal": 40,
      "quote": "production-ready",
      "quoteConfirmedInTranscript": true,
      "example": "`field-a`, time index 0, slice index 0, appearance 0, camera 0",
      "requestedAppearance": "Grayscale",
      "images": [
        {
          "file": "artifacts/measured/independent-reference.png",
          "label": "Independent reference",
          "alt": "Reference",
          "source": "evidence:paraview-model-package-03/initial/reference/field-a/slice-0-0-appearance-0-camera-0.png",
          "sha256": "b81d5f3ef8a0da7a79562d887ce31949e11b791cbb3a2b5868c8794229ee550f"
        },
        {
          "file": "artifacts/measured/haiku-native.png",
          "label": "Haiku native",
          "alt": "Haiku native",
          "source": "evidence:ai-paraview-authoring-claude-02/haiku-native/initial/workspace/submission/field-a/slice-0-0-appearance-0-camera-0.png",
          "sha256": "85824a92c59d3ee3cbd32f03613a9fff8751056133097c471417fc52fd015557"
        },
        {
          "file": "artifacts/measured/sonnet-native.png",
          "label": "Sonnet native",
          "alt": "Sonnet native",
          "source": "evidence:ai-paraview-authoring-claude-02/sonnet-native/initial/workspace/submission/field-a/slice-0-0-appearance-0-camera-0.png",
          "sha256": "791aba3297089ec58b433b97fd423078f91c3b94c181fa5f72c5c89ef5b56bd4"
        },
        {
          "file": "artifacts/measured/haiku-with-schemaxis.png",
          "label": "Haiku with Schemaxis",
          "alt": "Haiku Schemaxis",
          "source": "evidence:ai-paraview-authoring-claude-02/haiku-schemaxis/initial/workspace/submission/field-a/slice-0-0-appearance-0-camera-0.png",
          "sha256": "10fdc5a11cecd818509a40c117723b3b2c309dc3c29ab4277663c1feb0306f53"
        }
      ]
    },
    "frontier": {
      "astraNative": {
        "checks": "110/110",
        "seconds": "177.2",
        "secondsWhole": "177",
        "cost": "0.8828",
        "costShort": "0.88"
      },
      "astraSchemaxis": {
        "checks": "110/110",
        "seconds": "305.0",
        "secondsWhole": "305",
        "cost": "2.4853",
        "costShort": "2.49"
      },
      "opusSchemaxis": {
        "checks": "110/110",
        "seconds": "227.8",
        "secondsWhole": "228",
        "cost": "1.1867",
        "costShort": "1.19"
      },
      "preparedSlowerForEveryClaudeModel": true,
      "opusPrepared": {
        "schemaxis": "151",
        "native": "127"
      },
      "haikuPrepared": {
        "schemaxis": "129",
        "native": "79"
      }
    },
    "lunaVersusAstra": {
      "costReduction": 0.907715507866309,
      "timeIncrease": 0.6058181701528613,
      "display": {
        "costReductionApprox": "91",
        "costReduction": "90.8",
        "timeIncreaseApprox": "61",
        "timeIncrease": "60.6",
        "lunaCost": "0.0815",
        "astraCost": "0.8828",
        "lunaSeconds": "284.6",
        "astraSeconds": "177.2"
      },
      "conversationsPerCell": 1
    }
  },
  "sources": [
    {
      "file": "verification/2026-09-29-smaller-model-results.md",
      "sha256": "b256090172ee64880232f0898bf75e992c2fbaa7f2943603f03cf3262460e080"
    },
    {
      "file": "evidence/model-comparison-results.json",
      "sha256": "3176d4cafd420e6bc6475889a249d688c3a76edefb7f5ba4aa594282ff4b0ce8"
    },
    {
      "file": "evidence/claude-model-comparison-results.json",
      "sha256": "e1fa83404b5a2645f84d9b312882b09aacb6dcd5f8577ca65f05ebc5585701e7"
    },
    {
      "file": "evidence/openai-model-comparison-results.json",
      "sha256": "86836bccb2aaf0a1b28164eac392860ce27291f4d0cca18d7f861f272d70fb46"
    },
    {
      "file": "evidence/model-comparison-final-totals.json",
      "sha256": "2fc6f022175dd01b661476312daa0469db607bf470da721b5d74015678c5ade6"
    },
    {
      "file": "evidence/model-comparison-final-integrity.json",
      "sha256": "103e2896a26746766e45e1c8cd71c1d0b40c0aa152e717f392b9bc1de6ae946c"
    },
    {
      "file": "evidence/model-pricing-2026-09-29.json",
      "sha256": "a8b475838ba695295e10ca5d83e833767e65704b7f52dbe02c7a23057bdc033d"
    },
    {
      "file": "evidence/paraview-model-display-probe/report.json",
      "sha256": "5e223522ba6c5891a2da46cfc7eab32c3c45e8ecb34098044e92085888ae6e77"
    },
    {
      "file": "evidence/ai-paraview-authoring-claude-02/report.json",
      "sha256": "519cae371f830260fe0f923826573fb26e6f8ce59c44760202437c24c3986524"
    },
    {
      "file": "evidence/ai-paraview-authoring-codex-01/report.json",
      "sha256": "3dc6be073497bd5d2f08290eac6ef6d6a43156a61096b26c9167c955c22ea051"
    }
  ],
  "discrepancies": []
}