{
  "schemaVersion": 1,
  "archivedAt": "2026-09-05",
  "product": "Miaoclaw",
  "productVersion": "3.2.3",
  "evaluation": "WorkBuddy Bench task set maintainer reproduction",
  "officialRanking": false,
  "model": { "service": "Kimi Code", "modelId": "kimi-for-coding", "documentedMapping": "Kimi K2.7 Code", "documentation": "https://www.kimi.com/code/docs/kimi-code/models.html" },
  "totalTasks": 128,
  "answeredTasks": 127,
  "scoreScale": 100,
  "categories": [
    { "name": "office", "tasks": 50, "answered": 49, "missingInputCountedAsZero": 1, "meanScore": 54.9622, "answeredOnlyMeanScore": 56.0839 },
    { "name": "webpage_creation", "tasks": 70, "meanScore": 54.2857, "fullCredit": 21, "partialCredit": 30, "zeroCredit": 19, "measuresSearchAccuracy": false },
    { "name": "code_repair", "tasks": 8, "originalMeanScore": 25, "scoringEnvironmentCorrectedMeanScore": 84.6154, "samePatches": true, "newAnswering": false, "correctedFullCredit": 6, "correctedPartialCredit": 1, "correctedZeroCredit": 1 }
  ],
  "limitations": ["No same-condition competitor comparison", "Historical 3.2.3 results, not 3.3.1 acceptance", "Categories are not combined into an overall capability score", "Search, image answers and long-term memory need separate evaluation"]
}