{
  "baseBenchmark": "/Users/federico/ws/unleash-ai/docs/reports/2026-10-05-full-model-comparison",
  "scope": "Only prior raw fail cases with reviewed application defects, on the variant where they failed. Passing, partial and blocked cases excluded. Claude C98 excluded as obsolete wording expectation. New FAQ operational cases not run in this failure-only batch.",
  "models": [
    "openai/gpt-5.4-mini",
    "anthropic/claude-sonnet-4.6"
  ],
  "changedSourceSha256": {
    "packages/ai/components/workflows.ts": "227ac82593c9cc7c038ebbe8154a96aa01c49a33fd19dc90e6242077224787b5",
    "packages/ai/components/workflows.test.ts": "7f6d962dd353781c6be8277b68fb739d11d24a6d001d1bd0c90bfca5634bd36d",
    "apps/web/src/lib/aiResponsePolicy.ts": "eec1cf2dc79aadd8f7a56bcbb79e803315ff7aa0c13e41e92c92b10491ef8401",
    "apps/web/src/lib/aiResponsePolicy.test.ts": "5b5fb100acbb915063b3029aeb6e9e86c0159638f745d1209e4a7d9ba508e980",
    "apps/web/src/lib/aiFlows.ts": "6e9554011c948198eb520feea736b4e9dd2cd0380b8461099499acd952f16730",
    "apps/web/src/lib/aiFlows.test.ts": "c8a635711d44a031b1baaf450b3c3445c43755d45246859c97bc6a15255f37fb",
    "apps/web/src/lib/aiGuards.ts": "b4eb15946acfbff29be55c0812c571011a770e970a7887f657ba21b2aacc5c6b",
    "apps/web/src/lib/aiGuards.test.ts": "d10433160170232465b238bc97bc6bf5e0a7cd2fec64dc137ffd867c9445b8b8",
    "apps/web/src/lib/aiTelemetry.ts": "9e2f54458ebefbce870b8d00c77c0d9c1be3d0a2259054f893398b18b85cc37e",
    "apps/web/src/lib/aiTelemetry.test.ts": "3958d7c33334461541c5d46a9c4d98469aa6edff14e28f6a63c2d77edaf2d0a1",
    "apps/web/src/lib/handoffContact.ts": "fd67bca103cbeef32e72503f29cbac24f545655ba356bbabd5b28ad96d0bec1b",
    "apps/web/src/lib/handoffContact.test.ts": "3c565228ab73b9457d7458cdf631c05a90173057b5ef961e974f20e39151b408",
    "apps/web/src/routes/api/ai/index.ts": "80a0e70a4d3fd8aa34c502358b9b4ef0da58c1e123ccf26988e0760605c819ef",
    "packages/form-tests/tests/formBuilderConditions.test.ts": "258fc57f32568be03636269387a9436d0a93d98720aa347c8cb72987164260c8",
    "apps/web/scripts/faq-freshness.ts": "fc7b48c63b918e8985829d426491e8096e8453d18436b35e0ca618d55b91065e",
    "apps/web/scripts/faq-freshness-support.ts": "e993d8bc1b743e7dbb9e145070f5954f7eb59029c71f808310e2816f55d76b62",
    "apps/web/scripts/faq-freshness-support.test.ts": "e30a5ea9d46337ce1e34c976a3c3ad0b9bf2e56a253a5240a50c087d68ec6681",
    "packages/ai/components/api/pineconeUpload.ts": "b614ca58fe74957cf4a00f3bb022a3f2690c76ba78cecf3da9fc3df27b230e89",
    "apps/web/package.json": "211e06eeed24a4c653b5c140e57244d1c317c53c9e655ea40c98804032d131d9"
  },
  "noRealEnquiries": true,
  "liveSourcesNotFrozen": true,
  "isolatedSnapshotsNeedRefreshBeforeLaunch": false,
  "faqFreshnessOperationalCasesPerModel": 6,
  "faqFreshnessLiveRunStarted": false,
  "selection": {
    "openai": {
      "acceptance": 19,
      "e2e": [],
      "blocks": {
        "original-uat": [
          "C01",
          "C03",
          "C24",
          "C26",
          "C78",
          "C81",
          "C85"
        ],
        "uat-additions": [
          "C98",
          "C104",
          "C108",
          "C116"
        ],
        "regression": [
          "R01",
          "R02",
          "R07",
          "R14",
          "R26"
        ],
        "feedback": [
          "F04"
        ],
        "workflow-router": [
          "W06"
        ],
        "flow-selector": [
          "W06"
        ]
      }
    },
    "claude": {
      "acceptance": 17,
      "e2e": [
        "sponsorship-failure"
      ],
      "blocks": {
        "original-uat": [
          "C01",
          "C03",
          "C14",
          "C78",
          "C81"
        ],
        "uat-additions": [
          "C104",
          "C107",
          "C108",
          "C116"
        ],
        "regression": [
          "R01",
          "R02",
          "R07"
        ],
        "feedback": [
          "F03",
          "F04"
        ],
        "handoff-regression": [
          "H01"
        ],
        "workflow-router": [
          "W06"
        ],
        "flow-selector": [
          "W06"
        ]
      }
    }
  },
  "billingClearance": "User screenshot shows $40.21 Gateway credit and explicitly requests rerun of failed cases on 2026-10-05.",
  "actualCases": {
    "openai": 19,
    "claude": 18
  },
  "setupRecovery": "Suite wrapper corrected; stale baseline copy archived and excluded. Claude sponsorship-failure completed once before acceptance recovery; no duplicate."
}