{
  "format": "proofrun-repeatability-publication",
  "schemaVersion": 1,
  "contentHash": "sha256:ac15e88aba67131e7455e2763254ef979010e97a6a7f67ff1c5438d13c53043d",
  "preparedAt": "2026-08-13T18:10:33.485Z",
  "headline": "grok 4.6 ranged from 8 to 92 across five Prism Garden runs",
  "dek": "Five independent generations under one frozen condition produced a median craft score of 86; 4 met every requirement and 1 met none.",
  "study": {
    "id": "RP-MSROO2DW-E03530",
    "title": "Prism Garden · five-generation repeatability pilot",
    "status": "complete",
    "protocol": "proofrun-v0.7",
    "conditionHash": "sha256:da80b854ec27d19f4f6510213108896601e9e7c64603ecd10fe43314d1b3c404",
    "queueCommitment": "sha256:ae0cb263464d5d0a7ae46277474ed3580302c825c4830110a70d1685cd88d2b8",
    "createdAt": "2026-08-13T15:38:55.892Z",
    "revealedAt": "2026-08-13T16:08:09.942Z"
  },
  "test": {
    "id": "prism-garden",
    "title": "Prism Garden",
    "category": "Creative coding",
    "prompt": "Build a one-screen interactive web experience called PRISM GARDEN. Use a dark near-black background. Users click or drag to plant luminous geometric flowers; each flower should bloom with motion and slight variation. Include a small visible bloom counter and a Reset control. Make it polished, responsive, and immediately understandable. Use only HTML, CSS, and vanilla JavaScript in a single file. No external assets, fonts, or libraries. It must work offline. Spend time on the experience, not explanations.",
    "definitionHash": "curated:prism-garden:v2",
    "definitionVersion": 2,
    "executionMode": "web-single-file-v2",
    "requirements": [
      "Clicking plants a luminous geometric flower.",
      "Dragging plants multiple blooms without runaway creation.",
      "Blooms animate and show meaningful visual variation.",
      "The visible bloom counter remains accurate.",
      "Reset returns the experience to a clean initial state.",
      "The single-file experience remains responsive and works offline."
    ]
  },
  "condition": {
    "model": "x-ai/grok-4.6",
    "reasoningEffort": "high",
    "temperature": 0.7,
    "maxTokens": 16000,
    "wallLimitMinutes": 20,
    "writingGraceMinutes": 10,
    "generationCount": 5,
    "routing": {
      "mode": "auto",
      "endpointPinned": false,
      "fallbacksAllowed": true,
      "parametersRequired": true
    }
  },
  "method": {
    "targetCount": 5,
    "saltCount": 2,
    "duplicateCount": 2,
    "queueSize": 9,
    "lockedReviewCount": 9,
    "coldRescoreDelta": 6,
    "duplicateReviewDeltas": [
      {
        "originalLabel": "BLIND ARTIFACT 01",
        "duplicateLabel": "BLIND ARTIFACT 05",
        "delta": 0
      },
      {
        "originalLabel": "BLIND ARTIFACT 02",
        "duplicateLabel": "BLIND ARTIFACT 09",
        "delta": 0
      }
    ],
    "identitiesWithheldUntilAllReviewsLocked": true,
    "failuresRetainedWithoutReplacement": true
  },
  "summary": {
    "scores": [
      86,
      8,
      92,
      77,
      89
    ],
    "median": 86,
    "range": {
      "minimum": 8,
      "maximum": 92
    },
    "completedSubmissions": 5,
    "noSubmissionFailures": 0,
    "fullCompliance": 4,
    "complianceCollapses": 1,
    "totalReportedCostUsd": 0.43750599999999995,
    "totalWallTimeMs": 981177
  },
  "targets": [
    {
      "generationIndex": 1,
      "runId": "PR-MSROQWNW-25",
      "artifactOutcome": "completed_submission",
      "score": 86,
      "compliancePercent": 100,
      "requirements": [
        {
          "label": "Clicking plants a luminous geometric flower.",
          "judgment": "pass"
        },
        {
          "label": "Dragging plants multiple blooms without runaway creation.",
          "judgment": "pass"
        },
        {
          "label": "Blooms animate and show meaningful visual variation.",
          "judgment": "pass"
        },
        {
          "label": "The visible bloom counter remains accurate.",
          "judgment": "pass"
        },
        {
          "label": "Reset returns the experience to a clean initial state.",
          "judgment": "pass"
        },
        {
          "label": "The single-file experience remains responsive and works offline.",
          "judgment": "pass"
        }
      ],
      "craft": [
        {
          "label": "Visual design",
          "score": 7
        },
        {
          "label": "Interaction & usability",
          "score": 10
        },
        {
          "label": "Technical robustness",
          "score": 10
        },
        {
          "label": "Originality",
          "score": 6
        }
      ],
      "generation": {
        "provider": "xAI",
        "resolvedModel": "x-ai/grok-4.6",
        "finishReason": "stop",
        "wallTimeMs": 153714,
        "timeToFirstVisibleTokenMs": 69570,
        "reportedCostUsd": 0.062372,
        "promptTokens": 400,
        "completionTokens": 10294
      },
      "attempts": [
        {
          "ordinal": 1,
          "retryOfOrdinal": null,
          "lifecycle": "terminal",
          "artifactOutcome": "completed_submission",
          "terminalCause": "normal",
          "provider": "xAI",
          "startedAt": "2026-08-13T15:41:08.704Z",
          "finishedAt": "2026-08-13T15:43:42.418Z"
        }
      ],
      "evidence": {
        "state": "captured",
        "recipe": {
          "id": "prism-garden-standard-demonstration",
          "version": 1,
          "hash": "sha256:88ea9329a58b3f096bcdc31090ccf30352f726230f23472f830e9e7e5972c665",
          "status": "succeeded",
          "actionCount": 8,
          "passedActionCount": 8,
          "stabilizationMs": 1400,
          "assertions": [
            {
              "label": "Reset control detected and activated",
              "required": true,
              "status": "passed"
            },
            {
              "label": "Demonstration differs from the reset state",
              "required": true,
              "status": "passed"
            },
            {
              "label": "No new runtime errors during the recipe",
              "required": true,
              "status": "passed"
            }
          ],
          "failure": null
        },
        "image": {
          "id": "EV-bba1e7a4-c8e3-4a9d-8946-a7f2975e79c9",
          "path": "evidence/target-01.png",
          "contentHash": "sha256:d0c8d4c5c74e1b99a9c573fd682db397f331609038cd8fcea9f8870026a04bbd",
          "width": 1280,
          "height": 720,
          "capturedAt": "2026-08-13T16:59:34.469Z",
          "method": "browser-paint-evidence-recipe-v1",
          "label": "Standardized bloom demonstration",
          "caption": "Seven blooms planted at the frozen Prism Garden v1 coordinates after reset, then captured after the standardized stabilization window."
        }
      }
    },
    {
      "generationIndex": 2,
      "runId": "PR-MSROU7JW-67",
      "artifactOutcome": "completed_submission",
      "score": 8,
      "compliancePercent": 0,
      "requirements": [
        {
          "label": "Clicking plants a luminous geometric flower.",
          "judgment": "fail"
        },
        {
          "label": "Dragging plants multiple blooms without runaway creation.",
          "judgment": "fail"
        },
        {
          "label": "Blooms animate and show meaningful visual variation.",
          "judgment": "fail"
        },
        {
          "label": "The visible bloom counter remains accurate.",
          "judgment": "fail"
        },
        {
          "label": "Reset returns the experience to a clean initial state.",
          "judgment": "fail"
        },
        {
          "label": "The single-file experience remains responsive and works offline.",
          "judgment": "fail"
        }
      ],
      "craft": [
        {
          "label": "Visual design",
          "score": 2
        },
        {
          "label": "Interaction & usability",
          "score": 1
        },
        {
          "label": "Technical robustness",
          "score": 1
        },
        {
          "label": "Originality",
          "score": 5
        }
      ],
      "generation": {
        "provider": "xAI",
        "resolvedModel": "x-ai/grok-4.6",
        "finishReason": "stop",
        "wallTimeMs": 217367,
        "timeToFirstVisibleTokenMs": 136300,
        "reportedCostUsd": 0.094406,
        "promptTokens": 400,
        "completionTokens": 15633
      },
      "attempts": [
        {
          "ordinal": 1,
          "retryOfOrdinal": null,
          "lifecycle": "terminal",
          "artifactOutcome": "completed_submission",
          "terminalCause": "normal",
          "provider": "xAI",
          "startedAt": "2026-08-13T15:43:42.597Z",
          "finishedAt": "2026-08-13T15:47:19.964Z"
        }
      ],
      "evidence": {
        "state": "failed",
        "recipe": {
          "id": "prism-garden-standard-demonstration",
          "version": 1,
          "hash": "sha256:88ea9329a58b3f096bcdc31090ccf30352f726230f23472f830e9e7e5972c665",
          "status": "failed",
          "actionCount": 8,
          "passedActionCount": 8,
          "stabilizationMs": 1400,
          "assertions": [
            {
              "label": "Reset control detected and activated",
              "required": true,
              "status": "passed"
            },
            {
              "label": "Demonstration differs from the reset state",
              "required": true,
              "status": "failed"
            },
            {
              "label": "No new runtime errors during the recipe",
              "required": true,
              "status": "passed"
            }
          ],
          "failure": {
            "stage": "assertion",
            "message": "Demonstration differs from the reset state: The visual fingerprint did not change after the frozen actions."
          }
        },
        "image": null
      }
    },
    {
      "generationIndex": 3,
      "runId": "PR-MSROYVEP-BB",
      "artifactOutcome": "completed_submission",
      "score": 92,
      "compliancePercent": 100,
      "requirements": [
        {
          "label": "Clicking plants a luminous geometric flower.",
          "judgment": "pass"
        },
        {
          "label": "Dragging plants multiple blooms without runaway creation.",
          "judgment": "pass"
        },
        {
          "label": "Blooms animate and show meaningful visual variation.",
          "judgment": "pass"
        },
        {
          "label": "The visible bloom counter remains accurate.",
          "judgment": "pass"
        },
        {
          "label": "Reset returns the experience to a clean initial state.",
          "judgment": "pass"
        },
        {
          "label": "The single-file experience remains responsive and works offline.",
          "judgment": "pass"
        }
      ],
      "craft": [
        {
          "label": "Visual design",
          "score": 9
        },
        {
          "label": "Interaction & usability",
          "score": 10
        },
        {
          "label": "Technical robustness",
          "score": 9
        },
        {
          "label": "Originality",
          "score": 9
        }
      ],
      "generation": {
        "provider": "xAI",
        "resolvedModel": "x-ai/grok-4.6",
        "finishReason": "stop",
        "wallTimeMs": 214559,
        "timeToFirstVisibleTokenMs": 133826,
        "reportedCostUsd": 0.095006,
        "promptTokens": 400,
        "completionTokens": 15797
      },
      "attempts": [
        {
          "ordinal": 1,
          "retryOfOrdinal": null,
          "lifecycle": "terminal",
          "artifactOutcome": "completed_submission",
          "terminalCause": "normal",
          "provider": "xAI",
          "startedAt": "2026-08-13T15:47:20.141Z",
          "finishedAt": "2026-08-13T15:50:54.700Z"
        }
      ],
      "evidence": {
        "state": "captured",
        "recipe": {
          "id": "prism-garden-standard-demonstration",
          "version": 1,
          "hash": "sha256:88ea9329a58b3f096bcdc31090ccf30352f726230f23472f830e9e7e5972c665",
          "status": "succeeded",
          "actionCount": 8,
          "passedActionCount": 8,
          "stabilizationMs": 1400,
          "assertions": [
            {
              "label": "Reset control detected and activated",
              "required": true,
              "status": "passed"
            },
            {
              "label": "Demonstration differs from the reset state",
              "required": true,
              "status": "passed"
            },
            {
              "label": "No new runtime errors during the recipe",
              "required": true,
              "status": "passed"
            }
          ],
          "failure": null
        },
        "image": {
          "id": "EV-47f6196d-6bc4-40d8-b40c-fab2081645e4",
          "path": "evidence/target-03.png",
          "contentHash": "sha256:06569d2cabc2503b343ce5066f11e7ef130abcf49e996db2645ad236129bf996",
          "width": 1280,
          "height": 720,
          "capturedAt": "2026-08-13T16:59:40.685Z",
          "method": "browser-paint-evidence-recipe-v1",
          "label": "Standardized bloom demonstration",
          "caption": "Seven blooms planted at the frozen Prism Garden v1 coordinates after reset, then captured after the standardized stabilization window."
        }
      }
    },
    {
      "generationIndex": 4,
      "runId": "PR-MSRP3H67-EF",
      "artifactOutcome": "completed_submission",
      "score": 77,
      "compliancePercent": 100,
      "requirements": [
        {
          "label": "Clicking plants a luminous geometric flower.",
          "judgment": "pass"
        },
        {
          "label": "Dragging plants multiple blooms without runaway creation.",
          "judgment": "pass"
        },
        {
          "label": "Blooms animate and show meaningful visual variation.",
          "judgment": "pass"
        },
        {
          "label": "The visible bloom counter remains accurate.",
          "judgment": "pass"
        },
        {
          "label": "Reset returns the experience to a clean initial state.",
          "judgment": "pass"
        },
        {
          "label": "The single-file experience remains responsive and works offline.",
          "judgment": "pass"
        }
      ],
      "craft": [
        {
          "label": "Visual design",
          "score": 6
        },
        {
          "label": "Interaction & usability",
          "score": 9
        },
        {
          "label": "Technical robustness",
          "score": 9
        },
        {
          "label": "Originality",
          "score": 7
        }
      ],
      "generation": {
        "provider": "xAI",
        "resolvedModel": "x-ai/grok-4.6",
        "finishReason": "stop",
        "wallTimeMs": 160500,
        "timeToFirstVisibleTokenMs": 94823,
        "reportedCostUsd": 0.074576,
        "promptTokens": 400,
        "completionTokens": 12392
      },
      "attempts": [
        {
          "ordinal": 1,
          "retryOfOrdinal": null,
          "lifecycle": "terminal",
          "artifactOutcome": "completed_submission",
          "terminalCause": "normal",
          "provider": "xAI",
          "startedAt": "2026-08-13T15:50:54.996Z",
          "finishedAt": "2026-08-13T15:53:35.496Z"
        }
      ],
      "evidence": {
        "state": "captured",
        "recipe": {
          "id": "prism-garden-standard-demonstration",
          "version": 1,
          "hash": "sha256:88ea9329a58b3f096bcdc31090ccf30352f726230f23472f830e9e7e5972c665",
          "status": "succeeded",
          "actionCount": 8,
          "passedActionCount": 8,
          "stabilizationMs": 1400,
          "assertions": [
            {
              "label": "Reset control detected and activated",
              "required": true,
              "status": "passed"
            },
            {
              "label": "Demonstration differs from the reset state",
              "required": true,
              "status": "passed"
            },
            {
              "label": "No new runtime errors during the recipe",
              "required": true,
              "status": "passed"
            }
          ],
          "failure": null
        },
        "image": {
          "id": "EV-6a2787cd-a7ce-4e63-830a-955455fb5155",
          "path": "evidence/target-04.png",
          "contentHash": "sha256:20735c90a7235d912aaf9f5efc296ff0761c1423bc4ad2f1cb8772dfb181bf1e",
          "width": 1280,
          "height": 720,
          "capturedAt": "2026-08-13T16:59:43.853Z",
          "method": "browser-paint-evidence-recipe-v1",
          "label": "Standardized bloom demonstration",
          "caption": "Seven blooms planted at the frozen Prism Garden v1 coordinates after reset, then captured after the standardized stabilization window."
        }
      }
    },
    {
      "generationIndex": 5,
      "runId": "PR-MSRP6X96-8E",
      "artifactOutcome": "completed_submission",
      "score": 89,
      "compliancePercent": 100,
      "requirements": [
        {
          "label": "Clicking plants a luminous geometric flower.",
          "judgment": "pass"
        },
        {
          "label": "Dragging plants multiple blooms without runaway creation.",
          "judgment": "pass"
        },
        {
          "label": "Blooms animate and show meaningful visual variation.",
          "judgment": "pass"
        },
        {
          "label": "The visible bloom counter remains accurate.",
          "judgment": "pass"
        },
        {
          "label": "Reset returns the experience to a clean initial state.",
          "judgment": "pass"
        },
        {
          "label": "The single-file experience remains responsive and works offline.",
          "judgment": "pass"
        }
      ],
      "craft": [
        {
          "label": "Visual design",
          "score": 8
        },
        {
          "label": "Interaction & usability",
          "score": 10
        },
        {
          "label": "Technical robustness",
          "score": 10
        },
        {
          "label": "Originality",
          "score": 6
        }
      ],
      "generation": {
        "provider": "xAI",
        "resolvedModel": "x-ai/grok-4.6",
        "finishReason": "stop",
        "wallTimeMs": 235037,
        "timeToFirstVisibleTokenMs": 147031,
        "reportedCostUsd": 0.111146,
        "promptTokens": 400,
        "completionTokens": 18487
      },
      "attempts": [
        {
          "ordinal": 1,
          "retryOfOrdinal": null,
          "lifecycle": "terminal",
          "artifactOutcome": "completed_submission",
          "terminalCause": "normal",
          "provider": "xAI",
          "startedAt": "2026-08-13T15:53:35.831Z",
          "finishedAt": "2026-08-13T15:57:30.868Z"
        }
      ],
      "evidence": {
        "state": "captured",
        "recipe": {
          "id": "prism-garden-standard-demonstration",
          "version": 1,
          "hash": "sha256:88ea9329a58b3f096bcdc31090ccf30352f726230f23472f830e9e7e5972c665",
          "status": "succeeded",
          "actionCount": 8,
          "passedActionCount": 8,
          "stabilizationMs": 1400,
          "assertions": [
            {
              "label": "Reset control detected and activated",
              "required": true,
              "status": "passed"
            },
            {
              "label": "Demonstration differs from the reset state",
              "required": true,
              "status": "passed"
            },
            {
              "label": "No new runtime errors during the recipe",
              "required": true,
              "status": "passed"
            }
          ],
          "failure": null
        },
        "image": {
          "id": "EV-2f6b05bf-86b4-4ea4-a2f4-dc9acf93d69f",
          "path": "evidence/target-05.png",
          "contentHash": "sha256:6dc09ee72b54c5f41543673dca77ab3d76241444388f1597954e3da34a935995",
          "width": 1280,
          "height": 720,
          "capturedAt": "2026-08-13T16:59:47.047Z",
          "method": "browser-paint-evidence-recipe-v1",
          "label": "Standardized bloom demonstration",
          "caption": "Seven blooms planted at the frozen Prism Garden v1 coordinates after reset, then captured after the standardized stabilization window."
        }
      }
    }
  ],
  "disclosure": {
    "scope": "This record describes five independent generations from one model on one frozen test, route policy, temperature, token ceiling, and Evaluation v3 review session. It is not a universal model rating.",
    "limitations": [
      "Five generations are enough to expose material variability, not to estimate a stable population distribution.",
      "Two duplicate reviews measure consistency inside this review session only.",
      "A complete HTML submission can still fail every requirement; submission completion and requirement compliance are reported separately."
    ],
    "privateMaterialExcluded": [
      "researcher notebook",
      "blind queue secret",
      "raw model responses",
      "generated executable source",
      "credentials",
      "router metadata",
      "private reviewer notes"
    ],
    "generatedHtmlExecutedByPublicRecord": false
  }
}