{
  "schemaVersion": 1,
  "benchmarkVersion": "1.0.3",
  "exportedAt": "2026-07-15T02:55:59.254Z",
  "model": {
    "order": 8,
    "slug": "meta-llama-4-maverick",
    "company": "Meta",
    "model": "Llama 4 Maverick",
    "openRouterModelId": "meta-llama/llama-4-maverick",
    "expectedCanonicalSlug": "meta-llama/llama-4-maverick-17b-128e-instruct",
    "harnessModel": "openrouter/meta-llama/llama-4-maverick",
    "variant": null,
    "cohort": "former-reserve",
    "substitution": {
      "replaces": "Muse Spark 1.1",
      "reason": "Muse Spark 1.1 was not listed by the live OpenRouter Models API at the gateway freeze; Llama 4 Maverick is the strongest tool-capable Meta route currently listed."
    }
  },
  "route": "llama-4-maverick",
  "candidateCode": "C-1156",
  "record": {
    "runId": "meta-llama-4-maverick",
    "modelSlug": "meta-llama-4-maverick",
    "candidateCode": "C-1156",
    "runPosition": 2,
    "gateway": "OpenRouter",
    "requestedModelId": "meta-llama/llama-4-maverick",
    "expectedCanonicalSlug": "meta-llama/llama-4-maverick-17b-128e-instruct",
    "harnessModel": "openrouter/meta-llama/llama-4-maverick",
    "variant": null,
    "resolvedModelId": "meta-llama/llama-4-maverick",
    "resolvedCanonicalSlug": "meta-llama/llama-4-maverick-17b-128e-instruct",
    "startedAt": "2026-07-14T04:07:37.713Z",
    "finishedAt": "2026-07-14T04:07:51.096Z",
    "status": "completed",
    "promptSha256": "4fc686ce41533a72803ab11dd474f80c474666ea4154ead0fa0c90366042197f",
    "starterSha256": "6b745060ca29856fbeae84ea915909e83d6e7e7f05711aacf78918e2da5748a8",
    "elapsedMs": 13377,
    "inputTokens": 16216,
    "outputTokens": 219,
    "cachedTokens": null,
    "reportedCostUsd": 0.0034184,
    "exitCode": 0,
    "agentStepsObserved": 2,
    "interventions": 0
  },
  "evaluation": {
    "evaluatedAt": "2026-07-14T07:32:57.143Z",
    "modelSlug": "meta-llama-4-maverick",
    "installPassed": true,
    "buildPassed": true,
    "journeyPassed": false,
    "hardGatePassed": false,
    "automatedScore": 26,
    "automatedMaximum": 60,
    "passedChecks": [
      {
        "id": "F01",
        "title": "[F01|5][GATE] loads the correct substantial slice",
        "points": 5,
        "gate": true
      },
      {
        "id": "R01-mobile",
        "title": "[R01-mobile|3][GATE] no horizontal overflow at 360px",
        "points": 3,
        "gate": true
      },
      {
        "id": "R01-tablet",
        "title": "[R01-tablet|3][GATE] no horizontal overflow at 768px",
        "points": 3,
        "gate": true
      },
      {
        "id": "R01-desktop",
        "title": "[R01-desktop|2][GATE] no horizontal overflow at 1440px",
        "points": 2,
        "gate": true
      },
      {
        "id": "R03",
        "title": "[R03|2] images are local and load successfully when used",
        "points": 2,
        "gate": false
      },
      {
        "id": "A01",
        "title": "[A01|8][GATE] base page has no serious or critical axe violations",
        "points": 8,
        "gate": true
      },
      {
        "id": "A03",
        "title": "[A03|3] reduced motion suppresses long animations and transitions",
        "points": 3,
        "gate": false
      }
    ],
    "failedChecks": [
      {
        "id": "F02A",
        "title": "[F02A|6][GATE] filtering controls cards, URL state and direct loading",
        "points": 6,
        "gate": true
      },
      {
        "id": "F02B",
        "title": "[F02B|2] browser back and forward restore filter state",
        "points": 2,
        "gate": false
      },
      {
        "id": "F03",
        "title": "[F03|4][GATE] a keyboard-operable service button reveals supplied detail",
        "points": 4,
        "gate": true
      },
      {
        "id": "F04",
        "title": "[F04|4][GATE] dialog traps focus, closes with Escape and returns focus",
        "points": 4,
        "gate": true
      },
      {
        "id": "F05",
        "title": "[F05|4][GATE] the native-select form has errors and deterministic success",
        "points": 4,
        "gate": true
      },
      {
        "id": "R02",
        "title": "[R02|5] long copy is not clamped or clipped and missing media has a fallback",
        "points": 5,
        "gate": false
      },
      {
        "id": "R04",
        "title": "[R04|5] mobile navigation is present and keyboard usable",
        "points": 5,
        "gate": false
      },
      {
        "id": "A02",
        "title": "[A02|4][GATE] whole page remains accessible while the modal is open",
        "points": 4,
        "gate": true
      }
    ],
    "manualVisualScore": 8.25,
    "manualCodeScore": 0,
    "verificationScore": 0,
    "efficiencyScore": 5,
    "totalScore": 39.25,
    "capped": false,
    "validEvaluation": true,
    "infrastructureFailures": [],
    "evidenceWarnings": [],
    "manualReviews": [
      {
        "candidateCode": "C-1156",
        "reviewer": "R1",
        "visual": 7,
        "code": 0,
        "confidence": "high"
      },
      {
        "candidateCode": "C-1156",
        "reviewer": "R2",
        "visual": 9.5,
        "code": 0,
        "confidence": "high"
      }
    ],
    "thirdReviewRequired": false,
    "objectiveScoreEvidence": {
      "verification": {
        "ranCheck": false,
        "ranBuild": false,
        "progressTruthful": false
      },
      "efficiency": {
        "completedWithoutIntervention": true,
        "elapsedMs": 13377,
        "cleanExit": true,
        "reportedCostUsd": 0.0034184
      }
    },
    "rawScore": 39.25
  },
  "source": {
    "sha256": "6b745060ca29856fbeae84ea915909e83d6e7e7f05711aacf78918e2da5748a8",
    "archive": "/r0/downloads/llama-4-maverick-source.tar.gz",
    "archiveSha256": "2b577edd98358fb75ac3224987b53e83051eedf395f1d11c73bd1dbc44e45bdc",
    "exclusions": [
      "node_modules",
      "dist",
      ".git",
      ".opencode-state",
      "credential-bearing .env files"
    ]
  },
  "build": {
    "sha256": "26f0f87ca0ccbc4f1d9896fd717c6673848c8f534d353d683f4550e7a3f1b867",
    "publicPath": "/r0/llama-4-maverick/",
    "adaptations": [
      "npm run build -- --base=/r0/llama-4-maverick/",
      "Namespaced root /assets/ references in generated deployment files: assets/index-C03hvzRa.js"
    ]
  },
  "evidenceHashes": {
    "screenshots/360.png": "c2454f677e913700d4c7626909db6c2faf242ee4b9b95d23f7f3c07edb1833d6",
    "screenshots/768.png": "b9b8d6edaba034bef6462c81615add6f6d6c98c97b8c3a0dc0dae765f197b726",
    "screenshots/1440.png": "9e0f5073c241dbb07775f7b8e0f9881c0991c5249e02b9aaae13360416bb68cb",
    "interaction.webm": "186e4a92d80f86cf3be3d38bdd5c0a268be1a70899edadcf36da2d5c9f89917c",
    "diff.patch": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
  },
  "reviewDisclosure": "Visual and code scores were produced by two independent blind AI reviewer agents without model identities or access to each other’s scores."
}