{
  "evidence": {
    "runId": "8e9eb8e2-16b8-422b-8e59-5ed4b2c05d94",
    "evaluatedAt": "2026-09-27T21:19:58.394Z",
    "packageVersion": "0.8.1",
    "recipeFingerprint": "1d8a904b39336096ed678153226be54701c7248ebdf7f840f089be0b46f7c1fc",
    "datasetFingerprint": "c27c95c4f41f673b557b6d3f3af75b6cff20c1a9aff2e61bc0a160822be81981",
    "split": "held-out",
    "mode": "live",
    "policy": {
      "minConfidence": 0.8,
      "selectedOn": {
        "runId": "5e6a8dda-1b5b-4303-b7aa-67094d4dafe2",
        "datasetFingerprint": "56a4a143978b6c6b34d62be06f0e69c4bdc52001471bc90f35c231dbc18b2c55",
        "recipeFingerprint": "1d8a904b39336096ed678153226be54701c7248ebdf7f840f089be0b46f7c1fc",
        "model": "typesafe-ai/jev",
        "split": "development"
      }
    },
    "provenance": [
      {
        "method": "author-synthetic",
        "source": "AI-authored new scenario families for the September 2026 update; not independently human-reviewed. Previous inspected holdouts moved to development.",
        "cases": 40
      }
    ]
  },
  "recipe": "action-effects",
  "model": "typesafe-ai/jev",
  "models": [
    "typesafe-ai/jev"
  ],
  "cases": 40,
  "completed": 40,
  "failed": 0,
  "correct": 40,
  "wrong": 0,
  "ready": 20,
  "review": 20,
  "accuracy": 1,
  "completedAccuracy": 1,
  "accuracyInterval95": [
    0.9124,
    1
  ],
  "readyAccuracy": 1,
  "readyAccuracyInterval95": [
    0.8389,
    1
  ],
  "reviewRate": 0.5,
  "items": 200,
  "correctItems": 200,
  "itemAccuracy": 1,
  "contestedAccuracy": null,
  "adversarialAccuracy": null,
  "calibration": [
    {
      "confidence": "0-0.5",
      "cases": 0,
      "accuracy": null
    },
    {
      "confidence": "0.5-0.6",
      "cases": 0,
      "accuracy": null
    },
    {
      "confidence": "0.6-0.7",
      "cases": 16,
      "accuracy": 1
    },
    {
      "confidence": "0.7-0.8",
      "cases": 4,
      "accuracy": 1
    },
    {
      "confidence": "0.8-0.9",
      "cases": 10,
      "accuracy": 1
    },
    {
      "confidence": "0.9-1",
      "cases": 10,
      "accuracy": 1
    }
  ],
  "thresholdEvaluation": "frozen-policy",
  "thresholds": [],
  "suggestedMinConfidence": null,
  "latencyMs": {
    "p50": 261.06,
    "p95": 397.91,
    "totalCaseTime": 11371.99
  },
  "usage": {
    "requests": 40,
    "input_tokens": 37890,
    "output_tokens": 3840,
    "unknownResponses": 0,
    "failedRequests": 0,
    "models": [
      "typesafe-ai/jev"
    ]
  },
  "cost": {
    "model": "typesafe-ai/jev",
    "inputPerMillion": 0.042,
    "outputPerMillion": 0,
    "currency": "USD",
    "date": "2026-09-27",
    "source": "https://ai-gateway.vercel.sh/v1/models",
    "estimated": 0.00159138
  },
  "confusion": {
    "{\"labels.writesData.verdict\":\"absent\",\"labels.sendsMessage.verdict\":\"absent\",\"labels.spendsMoney.verdict\":\"absent\",\"labels.deletesData.verdict\":\"absent\",\"labels.callsExternal.verdict\":\"absent\"}": {
      "total": 20,
      "correct": 20,
      "misses": {}
    },
    "{\"labels.writesData.verdict\":\"absent\",\"labels.sendsMessage.verdict\":\"present\",\"labels.spendsMoney.verdict\":\"absent\",\"labels.deletesData.verdict\":\"absent\",\"labels.callsExternal.verdict\":\"present\"}": {
      "total": 10,
      "correct": 10,
      "misses": {}
    },
    "{\"labels.writesData.verdict\":\"present\",\"labels.sendsMessage.verdict\":\"absent\",\"labels.spendsMoney.verdict\":\"absent\",\"labels.deletesData.verdict\":\"absent\",\"labels.callsExternal.verdict\":\"absent\"}": {
      "total": 10,
      "correct": 10,
      "misses": {}
    }
  },
  "failures": [],
  "acceptance": {
    "minConfidence": 0.8,
    "minimumHeldOutReadyCases": 20,
    "minimumReadyAccuracy": 0.95,
    "maximumProviderFailures": 0,
    "met": true,
    "label": "Measured on these synthetic cases"
  },
  "gatewayIdentity": "Gateway returned the alias typesafe-ai/jev; backend version and BYOK credential use are not independently established."
}