[
  {
    "cell": "glm-openrouter",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.98,
        "order": {
          "a": "hermes",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_097873a303841d2f016a68849e692081a1a29e7cf1ec1b069d",
        "site_a_total": 81,
        "site_b_total": 97,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 31575,
          "input_tokens_details": {
            "cache_write_tokens": 31572,
            "cached_tokens": 0
          },
          "output_tokens": 4619,
          "output_tokens_details": {
            "reasoning_tokens": 3505
          },
          "total_tokens": 36194
        }
      },
      {
        "confidence": 0.97,
        "order": {
          "a": "pft",
          "b": "hermes"
        },
        "pass": "swapped",
        "response_id": "resp_099bf309fe88488e016a6884f8786c81a3a24b0edcd37c8f3f",
        "site_a_total": 95,
        "site_b_total": 80,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 31575,
          "input_tokens_details": {
            "cache_write_tokens": 31572,
            "cached_tokens": 0
          },
          "output_tokens": 5510,
          "output_tokens_details": {
            "reasoning_tokens": 4248
          },
          "total_tokens": 37085
        }
      }
    ],
    "verdict": "pft",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 1
  },
  {
    "cell": "glm-openrouter",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.95,
        "order": {
          "a": "hermes",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_02f4b76f1900fdae016a68856a5c14819dbc726b9f2afd45a9",
        "site_a_total": 94,
        "site_b_total": 85,
        "underlying_winner": "hermes",
        "usage": {
          "input_tokens": 34437,
          "input_tokens_details": {
            "cache_write_tokens": 34434,
            "cached_tokens": 0
          },
          "output_tokens": 4249,
          "output_tokens_details": {
            "reasoning_tokens": 3106
          },
          "total_tokens": 38686
        }
      },
      {
        "confidence": 0.9,
        "order": {
          "a": "pft",
          "b": "hermes"
        },
        "pass": "swapped",
        "response_id": "resp_04200e66a12eb497016a6885c466b0819e80e7e5c199ce0678",
        "site_a_total": 95,
        "site_b_total": 90,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 34437,
          "input_tokens_details": {
            "cache_write_tokens": 34434,
            "cached_tokens": 0
          },
          "output_tokens": 3586,
          "output_tokens_details": {
            "reasoning_tokens": 2378
          },
          "total_tokens": 38023
        }
      }
    ],
    "verdict": "tie_inconclusive_order_sensitive",
    "verdict_reason": "the balanced A/B orders did not select the same underlying site",
    "wave": 2
  },
  {
    "cell": "glm-openrouter",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.97,
        "order": {
          "a": "pft",
          "b": "hermes"
        },
        "pass": "normal",
        "response_id": "resp_06e7d8151aff76f5016a688607a2c881a28fa3c54ec89ee566",
        "site_a_total": 93,
        "site_b_total": 84,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 33141,
          "input_tokens_details": {
            "cache_write_tokens": 33138,
            "cached_tokens": 0
          },
          "output_tokens": 3572,
          "output_tokens_details": {
            "reasoning_tokens": 2588
          },
          "total_tokens": 36713
        }
      },
      {
        "confidence": 0.93,
        "order": {
          "a": "hermes",
          "b": "pft"
        },
        "pass": "swapped",
        "response_id": "resp_04f91257ce3a8bae016a68864ec60c819f8be4d52fb64e7db9",
        "site_a_total": 87,
        "site_b_total": 94,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 33141,
          "input_tokens_details": {
            "cache_write_tokens": 33138,
            "cached_tokens": 0
          },
          "output_tokens": 4768,
          "output_tokens_details": {
            "reasoning_tokens": 3587
          },
          "total_tokens": 37909
        }
      }
    ],
    "verdict": "pft",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 3
  },
  {
    "cell": "kimi-openrouter",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.97,
        "order": {
          "a": "hermes",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_0c2c99f3ba27bc1c016a6882d48dbc81a288f46c257b110001",
        "site_a_total": 90,
        "site_b_total": 98,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 29793,
          "input_tokens_details": {
            "cache_write_tokens": 29790,
            "cached_tokens": 0
          },
          "output_tokens": 3127,
          "output_tokens_details": {
            "reasoning_tokens": 2070
          },
          "total_tokens": 32920
        }
      },
      {
        "confidence": 0.95,
        "order": {
          "a": "pft",
          "b": "hermes"
        },
        "pass": "swapped",
        "response_id": "resp_058c66903f74b8f4016a688315b67c8191b610073acbeca3b8",
        "site_a_total": 97,
        "site_b_total": 90,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 29793,
          "input_tokens_details": {
            "cache_write_tokens": 29790,
            "cached_tokens": 0
          },
          "output_tokens": 3500,
          "output_tokens_details": {
            "reasoning_tokens": 2554
          },
          "total_tokens": 33293
        }
      }
    ],
    "verdict": "pft",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 1
  },
  {
    "cell": "kimi-openrouter",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.88,
        "order": {
          "a": "hermes",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_09bee529677f6e61016a68835a81a48191aa8a4ce383ca67cf",
        "site_a_total": 92,
        "site_b_total": 96,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 31467,
          "input_tokens_details": {
            "cache_write_tokens": 31464,
            "cached_tokens": 0
          },
          "output_tokens": 3654,
          "output_tokens_details": {
            "reasoning_tokens": 2342
          },
          "total_tokens": 35121
        }
      },
      {
        "confidence": 0.84,
        "order": {
          "a": "pft",
          "b": "hermes"
        },
        "pass": "swapped",
        "response_id": "resp_00e7c2a5b07dcf15016a6883a03ce881a2a41ce13291f428bb",
        "site_a_total": 96,
        "site_b_total": 92,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 31467,
          "input_tokens_details": {
            "cache_write_tokens": 31464,
            "cached_tokens": 0
          },
          "output_tokens": 4139,
          "output_tokens_details": {
            "reasoning_tokens": 2978
          },
          "total_tokens": 35606
        }
      }
    ],
    "verdict": "pft",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 2
  },
  {
    "cell": "kimi-openrouter",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.92,
        "order": {
          "a": "hermes",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_044eaa94865f9638016a6883efd628819c858f0184f826199c",
        "site_a_total": 90,
        "site_b_total": 94,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 31251,
          "input_tokens_details": {
            "cache_write_tokens": 31248,
            "cached_tokens": 0
          },
          "output_tokens": 4810,
          "output_tokens_details": {
            "reasoning_tokens": 3506
          },
          "total_tokens": 36061
        }
      },
      {
        "confidence": 0.9,
        "order": {
          "a": "pft",
          "b": "hermes"
        },
        "pass": "swapped",
        "response_id": "resp_09b50af7a87e7226016a688453036c81a09e7a41f9161c9aae",
        "site_a_total": 94,
        "site_b_total": 88,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 31251,
          "input_tokens_details": {
            "cache_write_tokens": 31248,
            "cached_tokens": 0
          },
          "output_tokens": 3739,
          "output_tokens_details": {
            "reasoning_tokens": 2433
          },
          "total_tokens": 34990
        }
      }
    ],
    "verdict": "pft",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 3
  },
  {
    "cell": "opus",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.95,
        "order": {
          "a": "cc",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_04a2bf03abdcc12c016a681d41e7ac819d93e2fadebc253660",
        "site_a_total": 94,
        "site_b_total": 90,
        "underlying_winner": "cc",
        "usage": {
          "input_tokens": 38703,
          "input_tokens_details": {
            "cache_write_tokens": 38700,
            "cached_tokens": 0
          },
          "output_tokens": 4085,
          "output_tokens_details": {
            "reasoning_tokens": 2921
          },
          "total_tokens": 42788
        }
      },
      {
        "confidence": 0.94,
        "order": {
          "a": "pft",
          "b": "cc"
        },
        "pass": "swapped",
        "response_id": "resp_0e996b0759b36842016a681d86fb60819f9fd673a6adb6cecb",
        "site_a_total": 91,
        "site_b_total": 97,
        "underlying_winner": "cc",
        "usage": {
          "input_tokens": 38703,
          "input_tokens_details": {
            "cache_write_tokens": 38700,
            "cached_tokens": 0
          },
          "output_tokens": 4167,
          "output_tokens_details": {
            "reasoning_tokens": 2995
          },
          "total_tokens": 42870
        }
      }
    ],
    "verdict": "cc",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 1
  },
  {
    "cell": "opus",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.94,
        "order": {
          "a": "cc",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_04949cbcc4b7c72b016a681dd1b4bc81a3957c183d81374832",
        "site_a_total": 89,
        "site_b_total": 95,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 38001,
          "input_tokens_details": {
            "cache_write_tokens": 37998,
            "cached_tokens": 0
          },
          "output_tokens": 4050,
          "output_tokens_details": {
            "reasoning_tokens": 3106
          },
          "total_tokens": 42051
        }
      },
      {
        "confidence": 0.96,
        "order": {
          "a": "pft",
          "b": "cc"
        },
        "pass": "swapped",
        "response_id": "resp_09ead2c455dafb03016a681e1d7dd0819d8e4316196190292d",
        "site_a_total": 95,
        "site_b_total": 88,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 38001,
          "input_tokens_details": {
            "cache_write_tokens": 37998,
            "cached_tokens": 0
          },
          "output_tokens": 3225,
          "output_tokens_details": {
            "reasoning_tokens": 1999
          },
          "total_tokens": 41226
        }
      }
    ],
    "verdict": "pft",
    "verdict_reason": "both balanced A/B orders selected the same underlying site",
    "wave": 2
  },
  {
    "cell": "opus",
    "model": "gpt-5.6-sol",
    "passes": [
      {
        "confidence": 0.96,
        "order": {
          "a": "cc",
          "b": "pft"
        },
        "pass": "normal",
        "response_id": "resp_08d806ead87241bc016a681e5c0cfc81a285668614d0e28ad3",
        "site_a_total": 97,
        "site_b_total": 97,
        "underlying_winner": "tie",
        "usage": {
          "input_tokens": 40971,
          "input_tokens_details": {
            "cache_write_tokens": 40968,
            "cached_tokens": 0
          },
          "output_tokens": 5534,
          "output_tokens_details": {
            "reasoning_tokens": 4469
          },
          "total_tokens": 46505
        }
      },
      {
        "confidence": 0.91,
        "order": {
          "a": "pft",
          "b": "cc"
        },
        "pass": "swapped",
        "response_id": "resp_0c64dac8f47e4d07016a681ebed364819296de2b4cbe22b939",
        "site_a_total": 96,
        "site_b_total": 93,
        "underlying_winner": "pft",
        "usage": {
          "input_tokens": 40971,
          "input_tokens_details": {
            "cache_write_tokens": 40968,
            "cached_tokens": 0
          },
          "output_tokens": 4691,
          "output_tokens_details": {
            "reasoning_tokens": 3403
          },
          "total_tokens": 45662
        }
      }
    ],
    "verdict": "tie_inconclusive_order_sensitive",
    "verdict_reason": "the balanced A/B orders did not select the same underlying site",
    "wave": 3
  }
]
