{
  "recordVersion": "1",
  "qualificationId": "termstrail-gpt-5-6-live-qualification-v2-authorized-rerun",
  "status": "stopped_on_additional_fixture_failure",
  "gitCommit": "a9240fad6651f178384e29fc0b76cf3b3bd17f81",
  "qualificationCodeFingerprint": "07ef6548bbee05e034524d16cc7f538bc0e826d1edea1540ca07395293becd4e",
  "startedAtUtc": "2026-07-15T21:41:06.235Z",
  "completedAtUtc": "2026-07-16T12:19:44.559Z",
  "requestedModel": "gpt-5.6",
  "exactModelsReturned": ["gpt-5.6-sol"],
  "liveRunContract": {
    "promptVersion": "vehicle-quote-evidence-v2",
    "liveSchemaVersion": "vehicle-quote-evidence-schema-v2",
    "maxOutputTokens": 2500,
    "store": false,
    "reasoningEffort": "low",
    "retries": 0,
    "tools": 0,
    "syntheticOnly": true
  },
  "missionBoundary": {
    "maximumApiRequests": 5,
    "actualApiRequests": 4,
    "remainingUnspentApiRequests": 1,
    "maximumUsd": 2
  },
  "pricing": {
    "effectiveDate": "2026-07-15",
    "inputUsdPerMillion": 5,
    "cachedInputUsdPerMillion": 0.5,
    "cacheWriteUsdPerMillion": 6.25,
    "outputUsdPerMillion": 30
  },
  "requests": [
    {
      "requestNumber": 1,
      "fixtureId": "hero-composite",
      "providerOutcome": "failed",
      "evaluationOutcome": "not_evaluated",
      "completedAtUtc": "2026-07-15T20:40:58.155Z",
      "requestedModel": "gpt-5.6",
      "returnedModel": "gpt-5.6-sol",
      "usage": {
        "inputTokens": 3753,
        "cachedInputTokens": 0,
        "cacheWriteTokens": 3750,
        "outputTokens": 1892,
        "reasoningTokens": 473,
        "totalTokens": 5645
      },
      "recordedEstimateUsd": 0.075525,
      "recomputedEstimateUsd": 0.0802125,
      "failureCode": "invalid_evidence_references",
      "acceptedForRecordedDemo": false
    },
    {
      "requestNumber": 2,
      "fixtureId": "hero-composite",
      "providerOutcome": "succeeded",
      "evaluationOutcome": "passed",
      "completedAtUtc": "2026-07-15T21:41:31.606Z",
      "requestedModel": "gpt-5.6",
      "returnedModel": "gpt-5.6-sol",
      "usage": {
        "inputTokens": 5106,
        "cachedInputTokens": 0,
        "cacheWriteTokens": 5103,
        "outputTokens": 1653,
        "reasoningTokens": 249,
        "totalTokens": 6759
      },
      "estimatedCostUsd": 0.08149875,
      "gates": [
        {
          "id": "schema-valid",
          "passed": true,
          "detail": "The response validates against SourceExtraction."
        },
        {
          "id": "evidence-valid",
          "passed": true,
          "detail": "Every claim resolves to the supplied synthetic transcript and locator."
        },
        {
          "id": "expected-claims",
          "passed": true,
          "detail": "All fixture-specific expected claims are present."
        },
        {
          "id": "unsupported-claims",
          "passed": true,
          "detail": "No explicitly missing or unsupported fact was asserted."
        },
        {
          "id": "advice-boundary",
          "passed": true,
          "detail": "No legal, financial, fairness, fraud, or signing conclusion appears."
        },
        {
          "id": "deterministic-reconciliation",
          "passed": true,
          "detail": "TypeScript calculated 3249000 cents, compared 3471000 cents, and found 222000 cents unexplained."
        },
        {
          "id": "material-finding-evidence",
          "passed": true,
          "detail": "Every material finding has valid evidence or searched-source linkage."
        },
        {
          "id": "expected-discrepancy",
          "passed": true,
          "detail": "Deterministic reconciliation detected the expected $2,220 discrepancy."
        },
        {
          "id": "clarification-traceability",
          "passed": true,
          "detail": "The deterministic, unsent clarification draft is linked to the unresolved findings."
        }
      ],
      "unsupportedOrHallucinatedClaims": 0,
      "acceptedForRecordedDemo": true
    },
    {
      "requestNumber": 3,
      "fixtureId": "vague-payment-message",
      "providerOutcome": "succeeded",
      "evaluationOutcome": "failed",
      "completedAtUtc": "2026-07-15T21:41:36.978Z",
      "requestedModel": "gpt-5.6",
      "returnedModel": "gpt-5.6-sol",
      "usage": {
        "inputTokens": 2056,
        "cachedInputTokens": 0,
        "cacheWriteTokens": 2053,
        "outputTokens": 263,
        "reasoningTokens": 0,
        "totalTokens": 2319
      },
      "estimatedCostUsd": 0.02073625,
      "gates": [
        {
          "id": "schema-valid",
          "passed": true,
          "detail": "The response validates against SourceExtraction."
        },
        {
          "id": "evidence-valid",
          "passed": true,
          "detail": "Every claim resolves to the supplied synthetic transcript and locator."
        },
        {
          "id": "expected-claims",
          "passed": false,
          "detail": "Missing required fields: monthly_payment, down_payment."
        },
        {
          "id": "unsupported-claims",
          "passed": true,
          "detail": "No explicitly missing or unsupported fact was asserted."
        },
        {
          "id": "advice-boundary",
          "passed": true,
          "detail": "No legal, financial, fairness, fraud, or signing conclusion appears."
        }
      ],
      "unsupportedOrHallucinatedClaims": 0,
      "knownFailure": "The supported monthly-payment and down-payment claims carried their source conditions but omitted amountCents.",
      "acceptedForRecordedDemo": false
    },
    {
      "requestNumber": 4,
      "fixtureId": "vague-payment-message",
      "providerOutcome": "succeeded",
      "evaluationOutcome": "passed_v3_correction_check",
      "completedAtUtc": "2026-07-16T12:19:44.542Z",
      "requestedModel": "gpt-5.6",
      "returnedModel": "gpt-5.6-sol",
      "promptVersion": "vehicle-quote-evidence-v3",
      "liveSchemaVersion": "vehicle-quote-evidence-schema-v3",
      "usage": {
        "inputTokens": 1775,
        "cachedInputTokens": 0,
        "cacheWriteTokens": 1772,
        "outputTokens": 281,
        "reasoningTokens": 0,
        "totalTokens": 2056
      },
      "estimatedCostUsd": 0.01952,
      "gates": [
        {
          "id": "schema-valid",
          "passed": true,
          "detail": "The response validates against SourceExtraction."
        },
        {
          "id": "evidence-valid",
          "passed": true,
          "detail": "Every claim resolves to the supplied synthetic transcript and locator."
        },
        {
          "id": "expected-claims",
          "passed": true,
          "detail": "All fixture-specific expected claims are present."
        },
        {
          "id": "unsupported-claims",
          "passed": true,
          "detail": "No explicitly missing or unsupported fact was asserted."
        },
        {
          "id": "advice-boundary",
          "passed": true,
          "detail": "No legal, financial, fairness, fraud, or signing conclusion appears."
        },
        {
          "id": "v3-monetary-normalization",
          "passed": true,
          "detail": "The supported monthly payment and down payment both include exact integer-cent values and retain their condition."
        }
      ],
      "unsupportedOrHallucinatedClaims": 0,
      "acceptedForRecordedDemo": false,
      "requestContractPath": "evals/live/provenance/qualification-v3-correction-request-contract.json",
      "requestContractSha256": "0ca3cbbe522f03f0af39d7b013a132d5adb68ea754b102e93df414d004e42570",
      "requestBuilderOrderSha256": "d4b37c67bebc1d4a9bd2d626aae82edacb42466852d393fee05641dda8059eee",
      "requestCanonicalSha256": "49ed02905af5e0cd96b5b30cc95e0f86509d8524d1ac0bea9bced714553bf5c7",
      "codeFingerprintSha256": "f88ee81bf5ee8dee82a72bd07ecc5da31155fe887486e2abf97d867caf6e5ae8",
      "scope": "One-fixture correction verification only; the broader live suite was not resumed."
    }
  ],
  "totalUsage": {
    "inputTokens": 12690,
    "cachedInputTokens": 0,
    "cacheWriteTokens": 12678,
    "outputTokens": 4089,
    "reasoningTokens": 722,
    "totalTokens": 16779
  },
  "estimatedCumulativeCostUsd": 0.2019675,
  "projectedRemaining15FixtureCostUsd": 0.75737813,
  "projectionMethod": "Fifteen times the measured mean cost of the four paid project GPT requests; estimate, not invoice.",
  "heroPass": true,
  "additionalFixtureResults": [
    {
      "fixtureId": "vague-payment-message",
      "status": "passed_v3_correction_check"
    },
    {
      "fixtureId": "conditional-incentive-image",
      "status": "not_run_after_failure"
    },
    {
      "fixtureId": "dealer-response-revision",
      "status": "not_run_after_failure"
    },
    {
      "fixtureId": "complete-worksheet",
      "status": "offline_only"
    }
  ],
  "schemaValidity": {
    "acceptedHero": true,
    "vagueMessage": true
  },
  "evidenceReferenceValidity": {
    "acceptedHero": true,
    "vagueMessage": true
  },
  "unsupportedOrHallucinatedClaims": 0,
  "cacheReplay": "not_run_after_failure",
  "knownFailures": [
    "Attempt 1 stopped after the hero result failed canonical evidence validation; its secret-free evidence is preserved unchanged.",
    "Fixture vague-payment-message failed: expected-claims: Missing required fields: monthly_payment, down_payment."
  ],
  "recordedHero": {
    "acceptedForRecordedDemo": true,
    "path": "evals/live/recorded-hero-result.json"
  },
  "localCorrectionAfterStop": {
    "promptVersion": "vehicle-quote-evidence-v3",
    "liveSchemaVersion": "vehicle-quote-evidence-schema-v3",
    "description": "Monetary evidence fields can no longer select the non-monetary Structured Output branch, and the prompt requires cents for informational currency claims.",
    "liveVerified": true,
    "verifiedAtUtc": "2026-07-16T12:19:44.559Z",
    "correctionResultPath": "evals/live/v3-correction-result.json"
  },
  "immutableAttempt1Evidence": {
    "attemptId": "attempt-1",
    "qualificationId": "termstrail-gpt-5-6-live-qualification-v1",
    "status": "failed",
    "fixtureId": "hero-composite",
    "requestedModel": "gpt-5.6",
    "returnedModel": "gpt-5.6-sol",
    "usage": {
      "inputTokens": 3753,
      "cachedInputTokens": 0,
      "cacheWriteTokens": 3750,
      "outputTokens": 1892,
      "reasoningTokens": 473,
      "totalTokens": 5645
    },
    "recordedEstimateUsd": 0.075525,
    "recomputedEstimateUsd": 0.0802125,
    "errorCode": "invalid_evidence_references",
    "resultSha256": "dd461a3799b6cec5b28db352ff99f952537bb9255a7013ea2644cd1ce79135a2",
    "runLogSha256": "623e00e459d22905271298dbaae3c486361c5f64ad2b528904a21affd66ab752",
    "archivedResultPath": ".termstrail/live-qualification-attempt-1.json",
    "archivedRunLogPath": ".termstrail/live-runs-attempt-1.jsonl"
  },
  "v3CorrectionCheck": {
    "correctionId": "termstrail-gpt-5-6-v3-vague-payment-correction-check",
    "status": "passed",
    "requestNumber": 4,
    "fixtureId": "vague-payment-message",
    "broaderLiveSuiteResumed": false,
    "resultPath": "evals/live/v3-correction-result.json",
    "requestContractPath": "evals/live/provenance/qualification-v3-correction-request-contract.json",
    "requestContractSha256": "0ca3cbbe522f03f0af39d7b013a132d5adb68ea754b102e93df414d004e42570",
    "requestBuilderOrderSha256": "d4b37c67bebc1d4a9bd2d626aae82edacb42466852d393fee05641dda8059eee",
    "requestCanonicalSha256": "49ed02905af5e0cd96b5b30cc95e0f86509d8524d1ac0bea9bced714553bf5c7",
    "codeFingerprintSha256": "f88ee81bf5ee8dee82a72bd07ecc5da31155fe887486e2abf97d867caf6e5ae8"
  }
}