{
  "run_id": "jev-20260919-50x3-v1",
  "control": {
    "arm": "control",
    "repeat": null,
    "answers": 150,
    "passing": 114,
    "pass_rate_percent": 76,
    "calls": 150,
    "charged_cents": 18,
    "charged_usd": 0.18,
    "cost_per_passing_answer_usd": 0.0015789473684210526,
    "provider_usage_cost_usd": 0.2504200000000001,
    "provider_usage_cost_coverage": "150/150",
    "router_charged_usd": 0,
    "router_provider_cost_usd": 0,
    "p95_latency_seconds": 12.47475774999999,
    "policy_promotions": 0,
    "policy_promotion_rate_percent": 0,
    "fallback_answers": 0,
    "fallback_rate_percent": 0,
    "malformed_final_answers": 33,
    "failed_calls": 0,
    "lanes": {
      "fast": 0,
      "balanced": 0,
      "frontier": 150
    },
    "buckets": {
      "lookup": {
        "passing": 30,
        "total": 30
      },
      "rewriting": {
        "passing": 30,
        "total": 30
      },
      "tool-planning": {
        "passing": 27,
        "total": 30
      },
      "multi-step": {
        "passing": 0,
        "total": 30
      },
      "high-stakes": {
        "passing": 27,
        "total": 30
      }
    }
  },
  "treatment": {
    "arm": "treatment",
    "repeat": null,
    "answers": 150,
    "passing": 120,
    "pass_rate_percent": 80,
    "calls": 300,
    "charged_cents": 164,
    "charged_usd": 1.64,
    "cost_per_passing_answer_usd": 0.013666666666666666,
    "provider_usage_cost_usd": 0.16728710199999997,
    "provider_usage_cost_coverage": "300/300",
    "router_charged_usd": 1.5,
    "router_provider_cost_usd": 0.004885902000000003,
    "p95_latency_seconds": 20.068269458000024,
    "policy_promotions": 104,
    "policy_promotion_rate_percent": 69.33333333333333,
    "fallback_answers": 0,
    "fallback_rate_percent": 0,
    "malformed_final_answers": 16,
    "failed_calls": 0,
    "lanes": {
      "fast": 29,
      "balanced": 23,
      "frontier": 98
    },
    "buckets": {
      "lookup": {
        "passing": 27,
        "total": 30
      },
      "rewriting": {
        "passing": 30,
        "total": 30
      },
      "tool-planning": {
        "passing": 27,
        "total": 30
      },
      "multi-step": {
        "passing": 9,
        "total": 30
      },
      "high-stakes": {
        "passing": 27,
        "total": 30
      }
    }
  },
  "repeats": [
    {
      "repeat": 1,
      "control": {
        "arm": "control",
        "repeat": 1,
        "answers": 50,
        "passing": 38,
        "pass_rate_percent": 76,
        "calls": 50,
        "charged_cents": 12,
        "charged_usd": 0.12,
        "cost_per_passing_answer_usd": 0.003157894736842105,
        "provider_usage_cost_usd": 0.083415,
        "provider_usage_cost_coverage": "50/50",
        "router_charged_usd": 0,
        "router_provider_cost_usd": 0,
        "p95_latency_seconds": 12.846859333999994,
        "policy_promotions": 0,
        "policy_promotion_rate_percent": 0,
        "fallback_answers": 0,
        "fallback_rate_percent": 0,
        "malformed_final_answers": 11,
        "failed_calls": 0,
        "lanes": {
          "fast": 0,
          "balanced": 0,
          "frontier": 50
        },
        "buckets": {
          "lookup": {
            "passing": 10,
            "total": 10
          },
          "rewriting": {
            "passing": 10,
            "total": 10
          },
          "tool-planning": {
            "passing": 9,
            "total": 10
          },
          "multi-step": {
            "passing": 0,
            "total": 10
          },
          "high-stakes": {
            "passing": 9,
            "total": 10
          }
        }
      },
      "treatment": {
        "arm": "treatment",
        "repeat": 1,
        "answers": 50,
        "passing": 40,
        "pass_rate_percent": 80,
        "calls": 100,
        "charged_cents": 54,
        "charged_usd": 0.54,
        "cost_per_passing_answer_usd": 0.013500000000000002,
        "provider_usage_cost_usd": 0.05650013400000001,
        "provider_usage_cost_coverage": "100/100",
        "router_charged_usd": 0.5,
        "router_provider_cost_usd": 0.0016286339999999995,
        "p95_latency_seconds": 20.32356150000001,
        "policy_promotions": 34,
        "policy_promotion_rate_percent": 68,
        "fallback_answers": 0,
        "fallback_rate_percent": 0,
        "malformed_final_answers": 5,
        "failed_calls": 0,
        "lanes": {
          "fast": 10,
          "balanced": 7,
          "frontier": 33
        },
        "buckets": {
          "lookup": {
            "passing": 9,
            "total": 10
          },
          "rewriting": {
            "passing": 10,
            "total": 10
          },
          "tool-planning": {
            "passing": 9,
            "total": 10
          },
          "multi-step": {
            "passing": 3,
            "total": 10
          },
          "high-stakes": {
            "passing": 9,
            "total": 10
          }
        }
      }
    },
    {
      "repeat": 2,
      "control": {
        "arm": "control",
        "repeat": 2,
        "answers": 50,
        "passing": 38,
        "pass_rate_percent": 76,
        "calls": 50,
        "charged_cents": 3,
        "charged_usd": 0.03,
        "cost_per_passing_answer_usd": 0.0007894736842105263,
        "provider_usage_cost_usd": 0.08369000000000001,
        "provider_usage_cost_coverage": "50/50",
        "router_charged_usd": 0,
        "router_provider_cost_usd": 0,
        "p95_latency_seconds": 12.006624542000004,
        "policy_promotions": 0,
        "policy_promotion_rate_percent": 0,
        "fallback_answers": 0,
        "fallback_rate_percent": 0,
        "malformed_final_answers": 11,
        "failed_calls": 0,
        "lanes": {
          "fast": 0,
          "balanced": 0,
          "frontier": 50
        },
        "buckets": {
          "lookup": {
            "passing": 10,
            "total": 10
          },
          "rewriting": {
            "passing": 10,
            "total": 10
          },
          "tool-planning": {
            "passing": 9,
            "total": 10
          },
          "multi-step": {
            "passing": 0,
            "total": 10
          },
          "high-stakes": {
            "passing": 9,
            "total": 10
          }
        }
      },
      "treatment": {
        "arm": "treatment",
        "repeat": 2,
        "answers": 50,
        "passing": 40,
        "pass_rate_percent": 80,
        "calls": 100,
        "charged_cents": 56,
        "charged_usd": 0.56,
        "cost_per_passing_answer_usd": 0.014000000000000002,
        "provider_usage_cost_usd": 0.05421473400000001,
        "provider_usage_cost_coverage": "100/100",
        "router_charged_usd": 0.5,
        "router_provider_cost_usd": 0.0016286339999999995,
        "p95_latency_seconds": 20.068269458000024,
        "policy_promotions": 34,
        "policy_promotion_rate_percent": 68,
        "fallback_answers": 0,
        "fallback_rate_percent": 0,
        "malformed_final_answers": 5,
        "failed_calls": 0,
        "lanes": {
          "fast": 10,
          "balanced": 8,
          "frontier": 32
        },
        "buckets": {
          "lookup": {
            "passing": 9,
            "total": 10
          },
          "rewriting": {
            "passing": 10,
            "total": 10
          },
          "tool-planning": {
            "passing": 9,
            "total": 10
          },
          "multi-step": {
            "passing": 3,
            "total": 10
          },
          "high-stakes": {
            "passing": 9,
            "total": 10
          }
        }
      }
    },
    {
      "repeat": 3,
      "control": {
        "arm": "control",
        "repeat": 3,
        "answers": 50,
        "passing": 38,
        "pass_rate_percent": 76,
        "calls": 50,
        "charged_cents": 3,
        "charged_usd": 0.03,
        "cost_per_passing_answer_usd": 0.0007894736842105263,
        "provider_usage_cost_usd": 0.08331500000000001,
        "provider_usage_cost_coverage": "50/50",
        "router_charged_usd": 0,
        "router_provider_cost_usd": 0,
        "p95_latency_seconds": 12.606598291999893,
        "policy_promotions": 0,
        "policy_promotion_rate_percent": 0,
        "fallback_answers": 0,
        "fallback_rate_percent": 0,
        "malformed_final_answers": 11,
        "failed_calls": 0,
        "lanes": {
          "fast": 0,
          "balanced": 0,
          "frontier": 50
        },
        "buckets": {
          "lookup": {
            "passing": 10,
            "total": 10
          },
          "rewriting": {
            "passing": 10,
            "total": 10
          },
          "tool-planning": {
            "passing": 9,
            "total": 10
          },
          "multi-step": {
            "passing": 0,
            "total": 10
          },
          "high-stakes": {
            "passing": 9,
            "total": 10
          }
        }
      },
      "treatment": {
        "arm": "treatment",
        "repeat": 3,
        "answers": 50,
        "passing": 40,
        "pass_rate_percent": 80,
        "calls": 100,
        "charged_cents": 54,
        "charged_usd": 0.54,
        "cost_per_passing_answer_usd": 0.013500000000000002,
        "provider_usage_cost_usd": 0.056572234000000006,
        "provider_usage_cost_coverage": "100/100",
        "router_charged_usd": 0.5,
        "router_provider_cost_usd": 0.0016286339999999997,
        "p95_latency_seconds": 17.65240083399997,
        "policy_promotions": 36,
        "policy_promotion_rate_percent": 72,
        "fallback_answers": 0,
        "fallback_rate_percent": 0,
        "malformed_final_answers": 6,
        "failed_calls": 0,
        "lanes": {
          "fast": 9,
          "balanced": 8,
          "frontier": 33
        },
        "buckets": {
          "lookup": {
            "passing": 9,
            "total": 10
          },
          "rewriting": {
            "passing": 10,
            "total": 10
          },
          "tool-planning": {
            "passing": 9,
            "total": 10
          },
          "multi-step": {
            "passing": 3,
            "total": 10
          },
          "high-stakes": {
            "passing": 9,
            "total": 10
          }
        }
      }
    }
  ],
  "delta": {
    "charged_usd": 1.46,
    "pass_rate_percentage_points": 4,
    "p95_latency_seconds": 7.593511708000035,
    "provider_usage_cost_usd": -0.08313289800000012
  },
  "acceptance_rule_met": false
}
