{
  "status": "preliminary synthetic router experiment",
  "source_revision": "da3479f",
  "corpus_sha256": "2b277ecc9b1b5c8d91072a8e51a226c1487a08c0b6885122416c14689ee747a2",
  "cases": 45,
  "reps": 2,
  "arms": [
    {
      "id": "jev",
      "name": "Jev 1.13",
      "color": "#be542f",
      "run": "main-jev-v1",
      "slices": {
        "all": {
          "n": 90,
          "correct": 69,
          "valid": 90,
          "recall_n": 30,
          "missed_recall": 12,
          "unnecessary_recall": 9,
          "confusion": {
            "flow": {
              "flow": 27,
              "deliberate": 0,
              "recall": 9,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 24,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 11,
              "deliberate": 1,
              "recall": 18,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 26,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 2,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 0,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 2,
              "deliberate": 0,
              "recall": 8,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 20,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 6,
          "unnecessary_recall": 4,
          "confusion": {
            "flow": {
              "flow": 8,
              "deliberate": 0,
              "recall": 4,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 5,
              "deliberate": 1,
              "recall": 4,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 23,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 4,
          "unnecessary_recall": 3,
          "confusion": {
            "flow": {
              "flow": 9,
              "deliberate": 0,
              "recall": 3,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 4,
              "deliberate": 0,
              "recall": 6,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 219.514,
      "p95_ms": 359.214,
      "incident_affected": false,
      "estimated_usd_per_1000": 0.04233693333333333,
      "cost_usage_n": 90
    },
    {
      "id": "laya",
      "name": "Laya multilingual",
      "color": "#927aac",
      "run": "main-low-v1",
      "slices": {
        "all": {
          "n": 90,
          "correct": 36,
          "valid": 90,
          "recall_n": 30,
          "missed_recall": 30,
          "unnecessary_recall": 0,
          "confusion": {
            "flow": {
              "flow": 36,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 24,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 30,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 12,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 10,
          "unnecessary_recall": 0,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 8,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 10,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 12,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 10,
          "unnecessary_recall": 0,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 8,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 10,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 12,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 10,
          "unnecessary_recall": 0,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 8,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 10,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 2094.926,
      "p95_ms": 2870.833,
      "incident_affected": false,
      "estimated_usd_per_1000": null,
      "cost_usage_n": null
    },
    {
      "id": "laya-english",
      "name": "Laya English",
      "color": "#a191b2",
      "run": "main-laya-english-v1",
      "slices": {
        "all": {
          "n": 90,
          "correct": 29,
          "valid": 89,
          "recall_n": 30,
          "missed_recall": 26,
          "unnecessary_recall": 12,
          "confusion": {
            "flow": {
              "flow": 16,
              "deliberate": 14,
              "recall": 6,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 8,
              "deliberate": 9,
              "recall": 6,
              "no_decision": 1
            },
            "recall": {
              "flow": 18,
              "deliberate": 8,
              "recall": 4,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 11,
          "valid": 29,
          "recall_n": 10,
          "missed_recall": 10,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 6,
              "deliberate": 4,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 2,
              "deliberate": 5,
              "recall": 0,
              "no_decision": 1
            },
            "recall": {
              "flow": 6,
              "deliberate": 4,
              "recall": 0,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 8,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 6,
          "unnecessary_recall": 8,
          "confusion": {
            "flow": {
              "flow": 2,
              "deliberate": 6,
              "recall": 4,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 2,
              "deliberate": 2,
              "recall": 4,
              "no_decision": 0
            },
            "recall": {
              "flow": 6,
              "deliberate": 0,
              "recall": 4,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 10,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 10,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 8,
              "deliberate": 4,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 4,
              "deliberate": 2,
              "recall": 2,
              "no_decision": 0
            },
            "recall": {
              "flow": 6,
              "deliberate": 4,
              "recall": 0,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 5191.529,
      "p95_ms": 8985.034,
      "incident_affected": true,
      "estimated_usd_per_1000": null,
      "cost_usage_n": null
    },
    {
      "id": "luna-codex-low",
      "name": "Luna low",
      "color": "#427143",
      "run": "main-codex-low-v1",
      "slices": {
        "all": {
          "n": 90,
          "correct": 78,
          "valid": 90,
          "recall_n": 30,
          "missed_recall": 0,
          "unnecessary_recall": 12,
          "confusion": {
            "flow": {
              "flow": 24,
              "deliberate": 0,
              "recall": 12,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 24,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 0,
              "deliberate": 0,
              "recall": 30,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 26,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 0,
          "unnecessary_recall": 4,
          "confusion": {
            "flow": {
              "flow": 8,
              "deliberate": 0,
              "recall": 4,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 0,
              "deliberate": 0,
              "recall": 10,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 26,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 0,
          "unnecessary_recall": 4,
          "confusion": {
            "flow": {
              "flow": 8,
              "deliberate": 0,
              "recall": 4,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 0,
              "deliberate": 0,
              "recall": 10,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 26,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 0,
          "unnecessary_recall": 4,
          "confusion": {
            "flow": {
              "flow": 8,
              "deliberate": 0,
              "recall": 4,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 0,
              "deliberate": 0,
              "recall": 10,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 1357.783,
      "p95_ms": 2917.678,
      "incident_affected": false,
      "estimated_usd_per_1000": 0.067878,
      "cost_usage_n": 90
    },
    {
      "id": "luna-codex-medium@main-codex-medium-v2",
      "name": "Luna medium · first",
      "color": "#678059",
      "run": "main-codex-medium-v2",
      "slices": {
        "all": {
          "n": 90,
          "correct": 78,
          "valid": 89,
          "recall_n": 30,
          "missed_recall": 4,
          "unnecessary_recall": 6,
          "confusion": {
            "flow": {
              "flow": 32,
              "deliberate": 1,
              "recall": 3,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 20,
              "recall": 3,
              "no_decision": 1
            },
            "recall": {
              "flow": 4,
              "deliberate": 0,
              "recall": 26,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 25,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 2,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 1,
              "recall": 1,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 7,
              "recall": 1,
              "no_decision": 0
            },
            "recall": {
              "flow": 2,
              "deliberate": 0,
              "recall": 8,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 28,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 0,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 0,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 0,
              "deliberate": 0,
              "recall": 10,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 25,
          "valid": 29,
          "recall_n": 10,
          "missed_recall": 2,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 5,
              "recall": 2,
              "no_decision": 1
            },
            "recall": {
              "flow": 2,
              "deliberate": 0,
              "recall": 8,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 2153.967,
      "p95_ms": 5505.593,
      "incident_affected": false,
      "estimated_usd_per_1000": 0.097226,
      "cost_usage_n": 89
    },
    {
      "id": "luna-codex-medium@main-codex-medium-v3",
      "name": "Luna medium · repeat",
      "color": "#89997b",
      "run": "main-codex-medium-v3",
      "slices": {
        "all": {
          "n": 90,
          "correct": 75,
          "valid": 90,
          "recall_n": 30,
          "missed_recall": 5,
          "unnecessary_recall": 10,
          "confusion": {
            "flow": {
              "flow": 30,
              "deliberate": 0,
              "recall": 6,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 20,
              "recall": 4,
              "no_decision": 0
            },
            "recall": {
              "flow": 5,
              "deliberate": 0,
              "recall": 25,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 27,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 1,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 0,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 8,
              "recall": 0,
              "no_decision": 0
            },
            "recall": {
              "flow": 1,
              "deliberate": 0,
              "recall": 9,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 24,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 2,
          "unnecessary_recall": 4,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 0,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 6,
              "recall": 2,
              "no_decision": 0
            },
            "recall": {
              "flow": 2,
              "deliberate": 0,
              "recall": 8,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 24,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 2,
          "unnecessary_recall": 4,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 0,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 6,
              "recall": 2,
              "no_decision": 0
            },
            "recall": {
              "flow": 2,
              "deliberate": 0,
              "recall": 8,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 2106.152,
      "p95_ms": 6041.613,
      "incident_affected": false,
      "estimated_usd_per_1000": 0.100161,
      "cost_usage_n": 90
    },
    {
      "id": "luna-codex-high",
      "name": "Luna high",
      "color": "#243e31",
      "run": "main-codex-high-v1",
      "slices": {
        "all": {
          "n": 90,
          "correct": 81,
          "valid": 90,
          "recall_n": 30,
          "missed_recall": 3,
          "unnecessary_recall": 6,
          "confusion": {
            "flow": {
              "flow": 33,
              "deliberate": 0,
              "recall": 3,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 21,
              "recall": 3,
              "no_decision": 0
            },
            "recall": {
              "flow": 3,
              "deliberate": 0,
              "recall": 27,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 29,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 0,
          "unnecessary_recall": 1,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 7,
              "recall": 1,
              "no_decision": 0
            },
            "recall": {
              "flow": 0,
              "deliberate": 0,
              "recall": 10,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 25,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 2,
          "unnecessary_recall": 3,
          "confusion": {
            "flow": {
              "flow": 10,
              "deliberate": 0,
              "recall": 2,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 7,
              "recall": 1,
              "no_decision": 0
            },
            "recall": {
              "flow": 2,
              "deliberate": 0,
              "recall": 8,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 27,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 1,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 11,
              "deliberate": 0,
              "recall": 1,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 7,
              "recall": 1,
              "no_decision": 0
            },
            "recall": {
              "flow": 1,
              "deliberate": 0,
              "recall": 9,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 2755.837,
      "p95_ms": 7119.028,
      "incident_affected": false,
      "estimated_usd_per_1000": 0.111356,
      "cost_usage_n": 90
    },
    {
      "id": "sol-codex-low",
      "name": "Sol low",
      "color": "#43798b",
      "run": "main-sol-low-v1",
      "slices": {
        "all": {
          "n": 90,
          "correct": 74,
          "valid": 89,
          "recall_n": 30,
          "missed_recall": 10,
          "unnecessary_recall": 5,
          "confusion": {
            "flow": {
              "flow": 36,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 18,
              "recall": 5,
              "no_decision": 1
            },
            "recall": {
              "flow": 10,
              "deliberate": 0,
              "recall": 20,
              "no_decision": 0
            }
          }
        },
        "en": {
          "n": 30,
          "correct": 24,
          "valid": 29,
          "recall_n": 10,
          "missed_recall": 3,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 5,
              "recall": 2,
              "no_decision": 1
            },
            "recall": {
              "flow": 3,
              "deliberate": 0,
              "recall": 7,
              "no_decision": 0
            }
          }
        },
        "uk": {
          "n": 30,
          "correct": 25,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 3,
          "unnecessary_recall": 2,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 6,
              "recall": 2,
              "no_decision": 0
            },
            "recall": {
              "flow": 3,
              "deliberate": 0,
              "recall": 7,
              "no_decision": 0
            }
          }
        },
        "mixed": {
          "n": 30,
          "correct": 25,
          "valid": 30,
          "recall_n": 10,
          "missed_recall": 4,
          "unnecessary_recall": 1,
          "confusion": {
            "flow": {
              "flow": 12,
              "deliberate": 0,
              "recall": 0,
              "no_decision": 0
            },
            "deliberate": {
              "flow": 0,
              "deliberate": 7,
              "recall": 1,
              "no_decision": 0
            },
            "recall": {
              "flow": 4,
              "deliberate": 0,
              "recall": 6,
              "no_decision": 0
            }
          }
        }
      },
      "median_ms": 3355.907,
      "p95_ms": 6557.053,
      "incident_affected": true,
      "estimated_usd_per_1000": 1.455618,
      "cost_usage_n": 89
    }
  ],
  "notes": [
    "Not a general model leaderboard.",
    "90 attempts per arm are 45 cases repeated twice; translations are correlated.",
    "All-attempt accuracy includes non-decisions. Recall miss denominator is 30 overall or 10 in each language.",
    "Latency is successful-decision roundtrip; runs sequential, not head-to-head.",
    "English Laya and Sol timings affected by client memory pressure; not a clean speed comparison.",
    "Costs are usage-based/API-equivalent estimates, not observed subscription charges. Unknown usage is not zero.",
    "Laya local hardware/power cost unknown, not plotted at zero.",
    "Both usable medium replicates retained; defective medium and historical OpenRouter runs excluded."
  ]
}