{
  "generated_at": "2026-11-02T14:40:55.932879+00:00",
  "model": "kimi-k3.1-pro",
  "scored_cases": 5315,
  "unique_prompts": 5311,
  "actual_response_calls": 5368,
  "composition": {
    "base": 4207,
    "jwcas": 30,
    "professional": 1078
  },
  "professional_evidence": {
    "direct": 1071,
    "exact_prompt_alias": 7,
    "missing": 0
  },
  "duplicate_note": "Four duplicate prompts remain separate source rows. The 60 InQA supplement samples are included within InQA+ 384, not added again.",
  "by_language": {
    "de": {
      "cases": 687,
      "objective_n": 686,
      "objective_correct": 521,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 1,
      "api_errors": 0,
      "objective_accuracy": 0.759475
    },
    "tr": {
      "cases": 4288,
      "objective_n": 4287,
      "objective_correct": 2575,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 1,
      "api_errors": 0,
      "objective_accuracy": 0.600653
    },
    "ja": {
      "cases": 340,
      "objective_n": 299,
      "objective_correct": 272,
      "heuristic_n": 10,
      "heuristic_pass": 10,
      "open_unscored": 31,
      "api_errors": 0,
      "objective_accuracy": 0.909699
    }
  },
  "by_dataset": {
    "MultiPragEval": {
      "cases": 300,
      "objective_n": 300,
      "objective_correct": 263,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.876667
    },
    "ImplicaTR": {
      "cases": 3888,
      "objective_n": 3888,
      "objective_correct": 2179,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.560442
    },
    "JaSocial": {
      "cases": 10,
      "objective_n": 0,
      "objective_correct": 0,
      "heuristic_n": 10,
      "heuristic_pass": 10,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": null
    },
    "Custom · objective": {
      "cases": 6,
      "objective_n": 6,
      "objective_correct": 6,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 1.0
    },
    "Custom · open": {
      "cases": 3,
      "objective_n": 0,
      "objective_correct": 0,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 3,
      "api_errors": 0,
      "objective_accuracy": null
    },
    "JWCAS": {
      "cases": 30,
      "objective_n": 0,
      "objective_correct": 0,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 30,
      "api_errors": 0,
      "objective_accuracy": null
    },
    "InQA+": {
      "cases": 384,
      "objective_n": 384,
      "objective_correct": 256,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.666667
    },
    "OpenMWE": {
      "cases": 297,
      "objective_n": 297,
      "objective_correct": 270,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.909091
    },
    "Turkish Idioms · paraphrase": {
      "cases": 200,
      "objective_n": 200,
      "objective_correct": 199,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.995
    },
    "Turkish Idioms · figurative": {
      "cases": 197,
      "objective_n": 197,
      "objective_correct": 195,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.989848
    }
  },
  "inqa_by_label": {
    "1": {
      "cases": 161,
      "objective_n": 161,
      "objective_correct": 118,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.732919
    },
    "2": {
      "cases": 90,
      "objective_n": 90,
      "objective_correct": 66,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.733333
    },
    "3": {
      "cases": 15,
      "objective_n": 15,
      "objective_correct": 7,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.466667
    },
    "4": {
      "cases": 69,
      "objective_n": 69,
      "objective_correct": 59,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.855072
    },
    "5": {
      "cases": 27,
      "objective_n": 27,
      "objective_correct": 6,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.222222
    },
    "6": {
      "cases": 22,
      "objective_n": 22,
      "objective_correct": 0,
      "heuristic_n": 0,
      "heuristic_pass": 0,
      "open_unscored": 0,
      "api_errors": 0,
      "objective_accuracy": 0.0
    }
  },
  "api_errors": 0,
  "cost_estimate": {
    "rates_usd_per_mtok": {
      "input": 2.0,
      "output": 10.0
    },
    "base_4207": {
      "previous_43_standard_usd": 0.116292,
      "first_1000_batch_60pct_usd": 0.7083732,
      "remaining_3164_direct_standard_usd": 3.876492,
      "combined_estimated_usd": 4.7011572
    },
    "supplement_90_direct": {
      "responses": 90,
      "prompt_tokens": 17015,
      "completion_tokens": 31486,
      "total_tokens": 48501,
      "standard_usd": 0.34889000000000003
    },
    "professional_1071_direct": {
      "responses": 1071,
      "prompt_tokens": 217209,
      "completion_tokens": 68519,
      "total_tokens": 285728,
      "standard_usd": 1.119608
    },
    "all_actual_responses": {
      "responses": 5368,
      "prompt_tokens": 1108772,
      "completion_tokens": 442436,
      "total_tokens": 1551208,
      "standard_usd": 6.641904
    },
    "canonical_5315_scored_responses": {
      "responses": 5315,
      "prompt_tokens": 1098732,
      "completion_tokens": 438406,
      "total_tokens": 1537138,
      "standard_usd": 6.581524
    },
    "combined_estimated_usd": 6.169655199999999,
    "note": "List-price estimate from returned usage, not an invoice. Includes earlier duplicate InQA sampling in actual execution; excludes the separate one-case Batch recovery smoke test. Taxes and separate cache-write charges are excluded; cache discounts are not inferred."
  }
}