{
  "study": "Seedream 5 Pro on anchor ladder v3",
  "session_id": "70869064-b007-4442-a6b2-a8364e416850",
  "data_revision": "2026-10-13",
  "snapshot_at": "2026-10-13T12:07:40.446046+00:00",
  "revision": {
    "date": "2026-10-13",
    "change": "First complete-assignment snapshot; accepted corrected joint interval, method D."
  },
  "baseline_version": "anchor_baseline_v3",
  "mub_commit": "db548d5745a4b687d9352dbb7145cbe2a102d487",
  "research_platform_commit": "ecbf32c7ceb8f3262df153110b294b6096459f9a",
  "ladder_card_sha256": "551b7207ac75d077d8e084b1be65678dd0cd0670e18ac4a5aa0e42037425efec",
  "baseline_source_sha256": "eea6d39dacf3216ffd2633234c2c058ccd7183b8bf26fb73b82d308d4ec69038",
  "ladder": {
    "model": {
      "name": "Seedream 5 Pro",
      "type": "T2I"
    },
    "headline": {
      "method": "D_corrected",
      "axis": "realism",
      "baseline_version": "anchor_baseline_v3",
      "mos_raw": 81.76794016996152,
      "mos": null,
      "ci_raw": [
        73.03280340826771,
        93.24333769187791
      ],
      "ci": [
        73.0,
        93.2
      ],
      "ci_halfwidth": 10.1052671418051,
      "correction_factor": 1.0954451150103321,
      "confidence": "low",
      "n_pairs": 288,
      "n_judgements": 863,
      "n_boot": 10000,
      "n_boot_ok": 10000,
      "seed": 12345,
      "failures": {},
      "connected": true,
      "degenerate": false,
      "separated": [],
      "verdict": "numeric",
      "equivalent_to_photo": null,
      "baseline_votes_digest": "15f830b594dd20708be42fd190d163a2e980ec3efc9d24f28707cc665274bb39",
      "model_votes_digest": "cb0cd2995d25d48377ef6a8c233219761ebb8985555a8d8750c017e0ef92a9f0",
      "coverage_note": "Approximate small-stratum correction; nominal coverage is not established.",
      "measured_mos": 81.8,
      "absolute_shown": false
    },
    "blocks": {},
    "interval_scope": {
      "n_raters": 13,
      "n_generations": 144,
      "n_scenes": 48,
      "n_categories": 8,
      "generations_per_scene_min": 3,
      "generations_per_scene_max": 3,
      "en": "The 95% interval quantifies scene-sampling uncertainty within 8 fixed MUB corpus categories, conditional on the observed panel (13 raters) and the recorded generations (3 per scene; 144 total). It does not estimate variability across arbitrary expert panels, other categories, or fresh generations."
    },
    "session_id": "70869064-b007-4442-a6b2-a8364e416850"
  },
  "coverage": {
    "session_id": "70869064-b007-4442-a6b2-a8364e416850",
    "session_status": "active",
    "assignment_status": "complete",
    "snapshot_at": "2026-10-13T12:07:40.446046+00:00",
    "evaluation_start_utc": "2026-09-30T05:59:56.365869",
    "evaluation_end_utc": "2026-10-13T09:48:53.426262",
    "created_at": "2026-09-30 05:19:04.726531+00:00",
    "planned_pair_judgments": 864,
    "valid_pair_judgments": 863,
    "axis_judgments": 2588,
    "explicit_skips": [
      {
        "pair_id": "10a27178-4522-47f2-85df-be99347a836a",
        "case_id": "layout_typography_04",
        "rater": "rater-04",
        "rater_role": "art_direction"
      }
    ],
    "missing_scale_answers": [
      {
        "pair_id": "73cc42f8-6897-4cf5-b923-4b7d076dbbcb",
        "case_id": "fashion_and_beauty_04",
        "rater": "rater-05",
        "axis": "prompt_adherence"
      }
    ],
    "axis_counts": {
      "aesthetics": 863,
      "prompt_adherence": 862,
      "believability": 863
    },
    "rater_roles": {
      "art_direction": 4,
      "content_distribution": 5,
      "quality_assurance": 4
    },
    "pair_coverage": {
      "3": 287,
      "2": 1
    },
    "scales": [
      "believability",
      "aesthetics",
      "prompt_adherence"
    ],
    "n_scenes": 48,
    "n_generations": 144,
    "n_pairs": 288,
    "n_raters": 13
  },
  "pairwise_analysis": {
    "n_boot": 2000,
    "seed": 12345,
    "unit": "case_id",
    "stratified_by": "eight categories, six scenes each",
    "quantiles": [
      0.025,
      0.975
    ],
    "scope": "conditional on the observed raters and recorded generations",
    "agreement": "production MUB Gwet AC1, complete three-rater pairs only"
  },
  "head_to_head": {
    "believability": {
      "Qwen Image 2.0": {
        "n": 432,
        "wins": 200,
        "ties": 122,
        "losses": 110,
        "win": {
          "share": 0.46296296296296297,
          "ci95": [
            0.4097222222222222,
            0.5139467592592589
          ]
        },
        "tie": {
          "share": 0.2824074074074074,
          "ci95": [
            0.24074074074074073,
            0.32407407407407407
          ]
        },
        "loss": {
          "share": 0.25462962962962965,
          "ci95": [
            0.21064814814814814,
            0.30092592592592593
          ]
        },
        "win_or_tie": {
          "share": 0.7453703703703703,
          "ci95": [
            0.6990740740740741,
            0.7893518518518519
          ]
        },
        "tie_adjusted": {
          "share": 0.6041666666666666,
          "ci95": [
            0.5601851851851852,
            0.6481770833333331
          ]
        },
        "win_minus_loss": {
          "share": 0.20833333333333334,
          "ci95": [
            0.12037037037037036,
            0.2963541666666663
          ]
        },
        "agreement": {
          "ac1": 0.09718936311157438,
          "n_complete_pairs": 144,
          "excluded_incomplete_pairs": 0,
          "ratings_per_item": 3
        }
      },
      "GPT Image 2": {
        "n": 431,
        "wins": 91,
        "ties": 136,
        "losses": 204,
        "win": {
          "share": 0.2111368909512761,
          "ci95": [
            0.15813953488372093,
            0.26451653806723135
          ]
        },
        "tie": {
          "share": 0.31554524361948955,
          "ci95": [
            0.2691415313225058,
            0.36430593151155777
          ]
        },
        "loss": {
          "share": 0.4733178654292343,
          "ci95": [
            0.4027777777777778,
            0.541860465116279
          ]
        },
        "win_or_tie": {
          "share": 0.5266821345707656,
          "ci95": [
            0.45813953488372094,
            0.5972222222222222
          ]
        },
        "tie_adjusted": {
          "share": 0.3689095127610209,
          "ci95": [
            0.3120469058821002,
            0.4271039244186045
          ]
        },
        "win_minus_loss": {
          "share": -0.26218097447795824,
          "ci95": [
            -0.37590618823579963,
            -0.14579215116279093
          ]
        },
        "agreement": {
          "ac1": 0.20208884755705073,
          "n_complete_pairs": 143,
          "excluded_incomplete_pairs": 1,
          "ratings_per_item": 3
        }
      }
    },
    "aesthetics": {
      "Qwen Image 2.0": {
        "n": 432,
        "wins": 284,
        "ties": 60,
        "losses": 88,
        "win": {
          "share": 0.6574074074074074,
          "ci95": [
            0.6041666666666666,
            0.7083333333333334
          ]
        },
        "tie": {
          "share": 0.1388888888888889,
          "ci95": [
            0.10185185185185185,
            0.17366898148148116
          ]
        },
        "loss": {
          "share": 0.2037037037037037,
          "ci95": [
            0.16435185185185186,
            0.24768518518518517
          ]
        },
        "win_or_tie": {
          "share": 0.7962962962962963,
          "ci95": [
            0.7523148148148148,
            0.8356481481481481
          ]
        },
        "tie_adjusted": {
          "share": 0.7268518518518519,
          "ci95": [
            0.6828703703703703,
            0.7685185185185185
          ]
        },
        "win_minus_loss": {
          "share": 0.4537037037037037,
          "ci95": [
            0.36574074074074076,
            0.5370370370370371
          ]
        },
        "agreement": {
          "ac1": 0.3363960032158036,
          "n_complete_pairs": 144,
          "excluded_incomplete_pairs": 0,
          "ratings_per_item": 3
        }
      },
      "GPT Image 2": {
        "n": 431,
        "wins": 129,
        "ties": 117,
        "losses": 185,
        "win": {
          "share": 0.29930394431554525,
          "ci95": [
            0.24074074074074073,
            0.35881710814642936
          ]
        },
        "tie": {
          "share": 0.271461716937355,
          "ci95": [
            0.2296983758700696,
            0.3109048723897912
          ]
        },
        "loss": {
          "share": 0.42923433874709976,
          "ci95": [
            0.3680402131782946,
            0.4906976744186046
          ]
        },
        "win_or_tie": {
          "share": 0.5707656612529002,
          "ci95": [
            0.5093023255813953,
            0.6319597868217053
          ]
        },
        "tie_adjusted": {
          "share": 0.43503480278422274,
          "ci95": [
            0.3793433335131927,
            0.4918609372470728
          ]
        },
        "win_minus_loss": {
          "share": -0.12993039443155452,
          "ci95": [
            -0.24131333297361462,
            -0.016278125505854425
          ]
        },
        "agreement": {
          "ac1": 0.1349760051619156,
          "n_complete_pairs": 143,
          "excluded_incomplete_pairs": 1,
          "ratings_per_item": 3
        }
      }
    },
    "prompt_adherence": {
      "Qwen Image 2.0": {
        "n": 432,
        "wins": 139,
        "ties": 241,
        "losses": 52,
        "win": {
          "share": 0.32175925925925924,
          "ci95": [
            0.2569444444444444,
            0.3889467592592589
          ]
        },
        "tie": {
          "share": 0.5578703703703703,
          "ci95": [
            0.4953125,
            0.6226851851851852
          ]
        },
        "loss": {
          "share": 0.12037037037037036,
          "ci95": [
            0.08796296296296297,
            0.1550925925925926
          ]
        },
        "win_or_tie": {
          "share": 0.8796296296296297,
          "ci95": [
            0.8449074074074074,
            0.9120370370370371
          ]
        },
        "tie_adjusted": {
          "share": 0.6006944444444444,
          "ci95": [
            0.5601851851851852,
            0.6423900462962961
          ]
        },
        "win_minus_loss": {
          "share": 0.2013888888888889,
          "ci95": [
            0.12037037037037036,
            0.28478009259259224
          ]
        },
        "agreement": {
          "ac1": 0.28413001912045943,
          "n_complete_pairs": 144,
          "excluded_incomplete_pairs": 0,
          "ratings_per_item": 3
        }
      },
      "GPT Image 2": {
        "n": 430,
        "wins": 62,
        "ties": 282,
        "losses": 86,
        "win": {
          "share": 0.14418604651162792,
          "ci95": [
            0.09578881009955777,
            0.19674513915364072
          ]
        },
        "tie": {
          "share": 0.6558139534883721,
          "ci95": [
            0.6004556074766355,
            0.7039799154334038
          ]
        },
        "loss": {
          "share": 0.2,
          "ci95": [
            0.14986543511133676,
            0.2517482517482518
          ]
        },
        "win_or_tie": {
          "share": 0.8,
          "ci95": [
            0.7482517482517482,
            0.8501345648886631
          ]
        },
        "tie_adjusted": {
          "share": 0.4720930232558139,
          "ci95": [
            0.4310577579851248,
            0.5163170163170163
          ]
        },
        "win_minus_loss": {
          "share": -0.05581395348837209,
          "ci95": [
            -0.1378844840297504,
            0.03263403263403263
          ]
        },
        "agreement": {
          "ac1": 0.4084556976984664,
          "n_complete_pairs": 142,
          "excluded_incomplete_pairs": 2,
          "ratings_per_item": 3
        }
      }
    }
  },
  "agreement": {
    "believability": {
      "ac1": 0.13196351183440747,
      "n_complete_pairs": 287,
      "excluded_incomplete_pairs": 1,
      "ratings_per_item": 3
    },
    "aesthetics": {
      "ac1": 0.21386728166095562,
      "n_complete_pairs": 287,
      "excluded_incomplete_pairs": 1,
      "ratings_per_item": 3
    },
    "prompt_adherence": {
      "ac1": 0.3419196588473342,
      "n_complete_pairs": 286,
      "excluded_incomplete_pairs": 2,
      "ratings_per_item": 3
    }
  },
  "by_category": {
    "Qwen Image 2.0": {
      "Fashion and beauty": {
        "n": 54,
        "wins": 14,
        "ties": 22,
        "losses": 18,
        "win": {
          "share": 0.25925925925925924,
          "ci95": [
            0.18518518518518517,
            0.35185185185185186
          ]
        },
        "tie": {
          "share": 0.4074074074074074,
          "ci95": [
            0.2962962962962963,
            0.5370370370370371
          ]
        },
        "loss": {
          "share": 0.3333333333333333,
          "ci95": [
            0.18518518518518517,
            0.46296296296296297
          ]
        },
        "win_or_tie": {
          "share": 0.6666666666666666,
          "ci95": [
            0.5370370370370371,
            0.8148148148148148
          ]
        },
        "tie_adjusted": {
          "share": 0.46296296296296297,
          "ci95": [
            0.37962962962962965,
            0.5833333333333334
          ]
        },
        "win_minus_loss": {
          "share": -0.07407407407407407,
          "ci95": [
            -0.24074074074074073,
            0.16666666666666666
          ]
        }
      },
      "Food and Beverage": {
        "n": 54,
        "wins": 28,
        "ties": 6,
        "losses": 20,
        "win": {
          "share": 0.5185185185185185,
          "ci95": [
            0.35185185185185186,
            0.6851851851851852
          ]
        },
        "tie": {
          "share": 0.1111111111111111,
          "ci95": [
            0.05555555555555555,
            0.16666666666666666
          ]
        },
        "loss": {
          "share": 0.37037037037037035,
          "ci95": [
            0.2037037037037037,
            0.5555555555555556
          ]
        },
        "win_or_tie": {
          "share": 0.6296296296296297,
          "ci95": [
            0.4444444444444444,
            0.7962962962962963
          ]
        },
        "tie_adjusted": {
          "share": 0.5740740740740741,
          "ci95": [
            0.39814814814814814,
            0.7407407407407407
          ]
        },
        "win_minus_loss": {
          "share": 0.14814814814814814,
          "ci95": [
            -0.2037037037037037,
            0.48148148148148145
          ]
        }
      },
      "Interior": {
        "n": 54,
        "wins": 26,
        "ties": 16,
        "losses": 12,
        "win": {
          "share": 0.48148148148148145,
          "ci95": [
            0.37037037037037035,
            0.5925925925925926
          ]
        },
        "tie": {
          "share": 0.2962962962962963,
          "ci95": [
            0.2222222222222222,
            0.37037037037037035
          ]
        },
        "loss": {
          "share": 0.2222222222222222,
          "ci95": [
            0.07407407407407407,
            0.35185185185185186
          ]
        },
        "win_or_tie": {
          "share": 0.7777777777777778,
          "ci95": [
            0.6481481481481481,
            0.9259259259259259
          ]
        },
        "tie_adjusted": {
          "share": 0.6296296296296297,
          "ci95": [
            0.5185185185185185,
            0.7407407407407407
          ]
        },
        "win_minus_loss": {
          "share": 0.25925925925925924,
          "ci95": [
            0.037037037037037035,
            0.48148148148148145
          ]
        }
      },
      "Layout & Typography": {
        "n": 54,
        "wins": 29,
        "ties": 17,
        "losses": 8,
        "win": {
          "share": 0.5370370370370371,
          "ci95": [
            0.3333333333333333,
            0.7037037037037037
          ]
        },
        "tie": {
          "share": 0.3148148148148148,
          "ci95": [
            0.2222222222222222,
            0.46296296296296297
          ]
        },
        "loss": {
          "share": 0.14814814814814814,
          "ci95": [
            0.07407407407407407,
            0.2222222222222222
          ]
        },
        "win_or_tie": {
          "share": 0.8518518518518519,
          "ci95": [
            0.7777777777777778,
            0.9259259259259259
          ]
        },
        "tie_adjusted": {
          "share": 0.6944444444444444,
          "ci95": [
            0.5648148148148148,
            0.8148148148148148
          ]
        },
        "win_minus_loss": {
          "share": 0.3888888888888889,
          "ci95": [
            0.12962962962962962,
            0.6296296296296297
          ]
        }
      },
      "People and Lifestyle": {
        "n": 54,
        "wins": 32,
        "ties": 18,
        "losses": 4,
        "win": {
          "share": 0.5925925925925926,
          "ci95": [
            0.42592592592592593,
            0.7597222222222197
          ]
        },
        "tie": {
          "share": 0.3333333333333333,
          "ci95": [
            0.14814814814814814,
            0.5370370370370371
          ]
        },
        "loss": {
          "share": 0.07407407407407407,
          "ci95": [
            0.018518518518518517,
            0.14814814814814814
          ]
        },
        "win_or_tie": {
          "share": 0.9259259259259259,
          "ci95": [
            0.8518518518518519,
            0.9814814814814815
          ]
        },
        "tie_adjusted": {
          "share": 0.7592592592592593,
          "ci95": [
            0.6574074074074074,
            0.8518518518518519
          ]
        },
        "win_minus_loss": {
          "share": 0.5185185185185185,
          "ci95": [
            0.3148148148148148,
            0.7037037037037037
          ]
        }
      },
      "Product and E-commerce": {
        "n": 54,
        "wins": 21,
        "ties": 18,
        "losses": 15,
        "win": {
          "share": 0.3888888888888889,
          "ci95": [
            0.2777777777777778,
            0.5185185185185185
          ]
        },
        "tie": {
          "share": 0.3333333333333333,
          "ci95": [
            0.2222222222222222,
            0.4444444444444444
          ]
        },
        "loss": {
          "share": 0.2777777777777778,
          "ci95": [
            0.12962962962962962,
            0.42592592592592593
          ]
        },
        "win_or_tie": {
          "share": 0.7222222222222222,
          "ci95": [
            0.5740740740740741,
            0.8703703703703703
          ]
        },
        "tie_adjusted": {
          "share": 0.5555555555555556,
          "ci95": [
            0.4444444444444444,
            0.6944444444444444
          ]
        },
        "win_minus_loss": {
          "share": 0.1111111111111111,
          "ci95": [
            -0.1111111111111111,
            0.3888888888888889
          ]
        }
      },
      "Social Marketing": {
        "n": 54,
        "wins": 26,
        "ties": 14,
        "losses": 14,
        "win": {
          "share": 0.48148148148148145,
          "ci95": [
            0.3333333333333333,
            0.6666666666666666
          ]
        },
        "tie": {
          "share": 0.25925925925925924,
          "ci95": [
            0.16666666666666666,
            0.35185185185185186
          ]
        },
        "loss": {
          "share": 0.25925925925925924,
          "ci95": [
            0.14814814814814814,
            0.3333333333333333
          ]
        },
        "win_or_tie": {
          "share": 0.7407407407407407,
          "ci95": [
            0.6666666666666666,
            0.8518518518518519
          ]
        },
        "tie_adjusted": {
          "share": 0.6111111111111112,
          "ci95": [
            0.5,
            0.7592592592592593
          ]
        },
        "win_minus_loss": {
          "share": 0.2222222222222222,
          "ci95": [
            0.0,
            0.5185185185185185
          ]
        }
      },
      "Street and Architecture": {
        "n": 54,
        "wins": 24,
        "ties": 11,
        "losses": 19,
        "win": {
          "share": 0.4444444444444444,
          "ci95": [
            0.35185185185185186,
            0.5555555555555556
          ]
        },
        "tie": {
          "share": 0.2037037037037037,
          "ci95": [
            0.1111111111111111,
            0.3148148148148148
          ]
        },
        "loss": {
          "share": 0.35185185185185186,
          "ci95": [
            0.25925925925925924,
            0.42592592592592593
          ]
        },
        "win_or_tie": {
          "share": 0.6481481481481481,
          "ci95": [
            0.5740740740740741,
            0.7407407407407407
          ]
        },
        "tie_adjusted": {
          "share": 0.5462962962962963,
          "ci95": [
            0.4722222222222222,
            0.6296296296296297
          ]
        },
        "win_minus_loss": {
          "share": 0.09259259259259259,
          "ci95": [
            -0.05555555555555555,
            0.25925925925925924
          ]
        }
      }
    },
    "GPT Image 2": {
      "Fashion and beauty": {
        "n": 54,
        "wins": 7,
        "ties": 14,
        "losses": 33,
        "win": {
          "share": 0.12962962962962962,
          "ci95": [
            0.037037037037037035,
            0.25925925925925924
          ]
        },
        "tie": {
          "share": 0.25925925925925924,
          "ci95": [
            0.12962962962962962,
            0.4074074074074074
          ]
        },
        "loss": {
          "share": 0.6111111111111112,
          "ci95": [
            0.4074074074074074,
            0.7962962962962963
          ]
        },
        "win_or_tie": {
          "share": 0.3888888888888889,
          "ci95": [
            0.2037037037037037,
            0.5925925925925926
          ]
        },
        "tie_adjusted": {
          "share": 0.25925925925925924,
          "ci95": [
            0.12962962962962962,
            0.42592592592592593
          ]
        },
        "win_minus_loss": {
          "share": -0.48148148148148145,
          "ci95": [
            -0.7407407407407407,
            -0.14814814814814814
          ]
        }
      },
      "Food and Beverage": {
        "n": 54,
        "wins": 9,
        "ties": 11,
        "losses": 34,
        "win": {
          "share": 0.16666666666666666,
          "ci95": [
            0.09212962962962964,
            0.25925925925925924
          ]
        },
        "tie": {
          "share": 0.2037037037037037,
          "ci95": [
            0.1111111111111111,
            0.2962962962962963
          ]
        },
        "loss": {
          "share": 0.6296296296296297,
          "ci95": [
            0.46296296296296297,
            0.7777777777777778
          ]
        },
        "win_or_tie": {
          "share": 0.37037037037037035,
          "ci95": [
            0.2222222222222222,
            0.5370370370370371
          ]
        },
        "tie_adjusted": {
          "share": 0.26851851851851855,
          "ci95": [
            0.1574074074074074,
            0.3888888888888889
          ]
        },
        "win_minus_loss": {
          "share": -0.46296296296296297,
          "ci95": [
            -0.6851851851851852,
            -0.2222222222222222
          ]
        }
      },
      "Interior": {
        "n": 54,
        "wins": 17,
        "ties": 24,
        "losses": 13,
        "win": {
          "share": 0.3148148148148148,
          "ci95": [
            0.16666666666666666,
            0.48148148148148145
          ]
        },
        "tie": {
          "share": 0.4444444444444444,
          "ci95": [
            0.25925925925925924,
            0.5935185185185134
          ]
        },
        "loss": {
          "share": 0.24074074074074073,
          "ci95": [
            0.07407407407407407,
            0.4074074074074074
          ]
        },
        "win_or_tie": {
          "share": 0.7592592592592593,
          "ci95": [
            0.5925925925925926,
            0.9259259259259259
          ]
        },
        "tie_adjusted": {
          "share": 0.5370370370370371,
          "ci95": [
            0.4074074074074074,
            0.6851851851851852
          ]
        },
        "win_minus_loss": {
          "share": 0.07407407407407407,
          "ci95": [
            -0.18518518518518517,
            0.37037037037037035
          ]
        }
      },
      "Layout & Typography": {
        "n": 53,
        "wins": 4,
        "ties": 19,
        "losses": 30,
        "win": {
          "share": 0.07547169811320754,
          "ci95": [
            0.0,
            0.15384615384615385
          ]
        },
        "tie": {
          "share": 0.3584905660377358,
          "ci95": [
            0.18867924528301888,
            0.5555555555555556
          ]
        },
        "loss": {
          "share": 0.5660377358490566,
          "ci95": [
            0.32075471698113206,
            0.7735849056603774
          ]
        },
        "win_or_tie": {
          "share": 0.4339622641509434,
          "ci95": [
            0.22641509433962265,
            0.6792452830188679
          ]
        },
        "tie_adjusted": {
          "share": 0.25471698113207547,
          "ci95": [
            0.11538461538461539,
            0.41509433962264153
          ]
        },
        "win_minus_loss": {
          "share": -0.49056603773584906,
          "ci95": [
            -0.7692307692307693,
            -0.16981132075471697
          ]
        }
      },
      "People and Lifestyle": {
        "n": 54,
        "wins": 17,
        "ties": 16,
        "losses": 21,
        "win": {
          "share": 0.3148148148148148,
          "ci95": [
            0.1111111111111111,
            0.556018518518516
          ]
        },
        "tie": {
          "share": 0.2962962962962963,
          "ci95": [
            0.16666666666666666,
            0.4074074074074074
          ]
        },
        "loss": {
          "share": 0.3888888888888889,
          "ci95": [
            0.2222222222222222,
            0.5555555555555556
          ]
        },
        "win_or_tie": {
          "share": 0.6111111111111112,
          "ci95": [
            0.4444444444444444,
            0.7777777777777778
          ]
        },
        "tie_adjusted": {
          "share": 0.46296296296296297,
          "ci95": [
            0.28703703703703703,
            0.6759259259259259
          ]
        },
        "win_minus_loss": {
          "share": -0.07407407407407407,
          "ci95": [
            -0.42592592592592593,
            0.35185185185185186
          ]
        }
      },
      "Product and E-commerce": {
        "n": 54,
        "wins": 12,
        "ties": 19,
        "losses": 23,
        "win": {
          "share": 0.2222222222222222,
          "ci95": [
            0.09259259259259259,
            0.4074074074074074
          ]
        },
        "tie": {
          "share": 0.35185185185185186,
          "ci95": [
            0.2222222222222222,
            0.46296296296296297
          ]
        },
        "loss": {
          "share": 0.42592592592592593,
          "ci95": [
            0.2037037037037037,
            0.6296296296296297
          ]
        },
        "win_or_tie": {
          "share": 0.5740740740740741,
          "ci95": [
            0.37037037037037035,
            0.7962962962962963
          ]
        },
        "tie_adjusted": {
          "share": 0.39814814814814814,
          "ci95": [
            0.24074074074074073,
            0.6018518518518519
          ]
        },
        "win_minus_loss": {
          "share": -0.2037037037037037,
          "ci95": [
            -0.5185185185185185,
            0.2037037037037037
          ]
        }
      },
      "Social Marketing": {
        "n": 54,
        "wins": 11,
        "ties": 15,
        "losses": 28,
        "win": {
          "share": 0.2037037037037037,
          "ci95": [
            0.07407407407407407,
            0.35185185185185186
          ]
        },
        "tie": {
          "share": 0.2777777777777778,
          "ci95": [
            0.14814814814814814,
            0.4074074074074074
          ]
        },
        "loss": {
          "share": 0.5185185185185185,
          "ci95": [
            0.25925925925925924,
            0.7407407407407407
          ]
        },
        "win_or_tie": {
          "share": 0.48148148148148145,
          "ci95": [
            0.25925925925925924,
            0.7407407407407407
          ]
        },
        "tie_adjusted": {
          "share": 0.3425925925925926,
          "ci95": [
            0.17592592592592593,
            0.5462962962962963
          ]
        },
        "win_minus_loss": {
          "share": -0.3148148148148148,
          "ci95": [
            -0.6481481481481481,
            0.09259259259259259
          ]
        }
      },
      "Street and Architecture": {
        "n": 54,
        "wins": 14,
        "ties": 18,
        "losses": 22,
        "win": {
          "share": 0.25925925925925924,
          "ci95": [
            0.12962962962962962,
            0.4074074074074074
          ]
        },
        "tie": {
          "share": 0.3333333333333333,
          "ci95": [
            0.24074074074074073,
            0.48148148148148145
          ]
        },
        "loss": {
          "share": 0.4074074074074074,
          "ci95": [
            0.24074074074074073,
            0.5925925925925926
          ]
        },
        "win_or_tie": {
          "share": 0.5925925925925926,
          "ci95": [
            0.4074074074074074,
            0.7592592592592593
          ]
        },
        "tie_adjusted": {
          "share": 0.42592592592592593,
          "ci95": [
            0.26851851851851855,
            0.5648148148148148
          ]
        },
        "win_minus_loss": {
          "share": -0.14814814814814814,
          "ci95": [
            -0.46296296296296297,
            0.12962962962962962
          ]
        }
      }
    }
  },
  "by_role": {
    "Qwen Image 2.0": {
      "art_direction": {
        "n": 133,
        "wins": 42,
        "ties": 59,
        "losses": 32,
        "win": {
          "share": 0.3157894736842105,
          "ci95": [
            0.23529411764705882,
            0.4015151515151515
          ]
        },
        "tie": {
          "share": 0.44360902255639095,
          "ci95": [
            0.36027113970588237,
            0.5303030303030303
          ]
        },
        "loss": {
          "share": 0.24060150375939848,
          "ci95": [
            0.16417910447761194,
            0.3203199546755725
          ]
        },
        "win_or_tie": {
          "share": 0.7593984962406015,
          "ci95": [
            0.6796800453244275,
            0.835820895522388
          ]
        },
        "tie_adjusted": {
          "share": 0.5375939849624061,
          "ci95": [
            0.4732773047563124,
            0.6037037037037037
          ]
        },
        "win_minus_loss": {
          "share": 0.07518796992481203,
          "ci95": [
            -0.05344539048737522,
            0.2074074074074074
          ]
        }
      },
      "content_distribution": {
        "n": 166,
        "wins": 73,
        "ties": 42,
        "losses": 51,
        "win": {
          "share": 0.4397590361445783,
          "ci95": [
            0.36645962732919257,
            0.5149792439660555
          ]
        },
        "tie": {
          "share": 0.25301204819277107,
          "ci95": [
            0.19252943088256536,
            0.31928334640336486
          ]
        },
        "loss": {
          "share": 0.3072289156626506,
          "ci95": [
            0.23780487804878048,
            0.3712574850299401
          ]
        },
        "win_or_tie": {
          "share": 0.6927710843373494,
          "ci95": [
            0.6287425149700598,
            0.7621951219512195
          ]
        },
        "tie_adjusted": {
          "share": 0.5662650602409639,
          "ci95": [
            0.5059880239520959,
            0.6317365269461078
          ]
        },
        "win_minus_loss": {
          "share": 0.13253012048192772,
          "ci95": [
            0.011976047904191617,
            0.2634730538922156
          ]
        }
      },
      "quality_assurance": {
        "n": 133,
        "wins": 85,
        "ties": 21,
        "losses": 27,
        "win": {
          "share": 0.6390977443609023,
          "ci95": [
            0.5648705016357689,
            0.7164179104477612
          ]
        },
        "tie": {
          "share": 0.15789473684210525,
          "ci95": [
            0.10368464052287582,
            0.21643381927571434
          ]
        },
        "loss": {
          "share": 0.20300751879699247,
          "ci95": [
            0.12878787878787878,
            0.2805755395683453
          ]
        },
        "win_or_tie": {
          "share": 0.7969924812030075,
          "ci95": [
            0.7194244604316546,
            0.8712121212121212
          ]
        },
        "tie_adjusted": {
          "share": 0.7180451127819549,
          "ci95": [
            0.6492537313432836,
            0.7868217054263565
          ]
        },
        "win_minus_loss": {
          "share": 0.43609022556390975,
          "ci95": [
            0.29850746268656714,
            0.5736434108527132
          ]
        }
      }
    },
    "GPT Image 2": {
      "art_direction": {
        "n": 132,
        "wins": 16,
        "ties": 59,
        "losses": 57,
        "win": {
          "share": 0.12121212121212122,
          "ci95": [
            0.06870229007633588,
            0.1838433257918551
          ]
        },
        "tie": {
          "share": 0.44696969696969696,
          "ci95": [
            0.37119107744107743,
            0.5234468005952381
          ]
        },
        "loss": {
          "share": 0.4318181818181818,
          "ci95": [
            0.34350030645790386,
            0.518796992481203
          ]
        },
        "win_or_tie": {
          "share": 0.5681818181818182,
          "ci95": [
            0.48120300751879697,
            0.6564996935420961
          ]
        },
        "tie_adjusted": {
          "share": 0.3446969696969697,
          "ci95": [
            0.28518446180555557,
            0.4111111111111111
          ]
        },
        "win_minus_loss": {
          "share": -0.3106060606060606,
          "ci95": [
            -0.4296310763888889,
            -0.17777777777777778
          ]
        }
      },
      "content_distribution": {
        "n": 166,
        "wins": 32,
        "ties": 51,
        "losses": 83,
        "win": {
          "share": 0.1927710843373494,
          "ci95": [
            0.12352759622367467,
            0.263817784359354
          ]
        },
        "tie": {
          "share": 0.3072289156626506,
          "ci95": [
            0.23214285714285715,
            0.381831550802139
          ]
        },
        "loss": {
          "share": 0.5,
          "ci95": [
            0.41277396053558846,
            0.5868263473053892
          ]
        },
        "win_or_tie": {
          "share": 0.5,
          "ci95": [
            0.41317365269461076,
            0.5872260394644115
          ]
        },
        "tie_adjusted": {
          "share": 0.3463855421686747,
          "ci95": [
            0.27744880264244426,
            0.4186868732414688
          ]
        },
        "win_minus_loss": {
          "share": -0.3072289156626506,
          "ci95": [
            -0.4451023947151115,
            -0.1626262535170624
          ]
        }
      },
      "quality_assurance": {
        "n": 133,
        "wins": 43,
        "ties": 26,
        "losses": 64,
        "win": {
          "share": 0.3233082706766917,
          "ci95": [
            0.22900763358778625,
            0.4253731343283582
          ]
        },
        "tie": {
          "share": 0.19548872180451127,
          "ci95": [
            0.1277957019414207,
            0.26619904076738604
          ]
        },
        "loss": {
          "share": 0.48120300751879697,
          "ci95": [
            0.38847847095457966,
            0.5725443421643465
          ]
        },
        "win_or_tie": {
          "share": 0.518796992481203,
          "ci95": [
            0.4274556578356533,
            0.6115215290454203
          ]
        },
        "tie_adjusted": {
          "share": 0.42105263157894735,
          "ci95": [
            0.3320610687022901,
            0.5113658049965302
          ]
        },
        "win_minus_loss": {
          "share": -0.15789473684210525,
          "ci95": [
            -0.33587786259541985,
            0.022731609993060353
          ]
        }
      }
    }
  },
  "rater_direction": {
    "Qwen Image 2.0": {
      "more_wins": 10,
      "equal": 0,
      "more_losses": 3
    },
    "GPT Image 2": {
      "more_wins": 1,
      "equal": 1,
      "more_losses": 11
    }
  },
  "side_balance": {
    "model_left": 437,
    "n": 863,
    "fraction": 0.5063731170336037
  },
  "generation_settings": {
    "endpoint": [
      "bytedance/seedream-v5.0-pro"
    ],
    "resolution": [
      "2k"
    ],
    "optimization": [
      "standard"
    ],
    "seed_supported": false,
    "median_wall_sec": 62.463,
    "n_native_dimensions": {
      "1672*2508": 63,
      "2508*1672": 69,
      "2364*1773": 3,
      "1831*2289": 6,
      "1773*2364": 3
    },
    "n_evaluation_dimensions": {
      "1365*2048": 63,
      "2048*1365": 69,
      "2048*1536": 3,
      "1638*2048": 6,
      "1536*2048": 3
    }
  },
  "image_assets": {
    "total": 240,
    "seedream": 144,
    "frozen_anchors": 96
  },
  "limits": [
    "Observed panel and recorded generations are fixed; intervals describe scene variation within eight fixed categories.",
    "Three Seedream outputs per scene are compared with one frozen image per anchor; anchor generation variance is not sampled.",
    "Ladder method D jointly resamples baseline and tested-model scenes within category, refits calibration and applies sqrt(6/5) about the replicate mean.",
    "The small-stratum correction is approximate; nominal coverage is not established.",
    "Ladder status is low: raw half-width exceeds 10. No public headline index is released.",
    "Low to modest agreement limits all scale-level findings to directional evidence for this panel.",
    "One explicit pair skip; one further missing prompt-adherence answer. Neither is converted into a tie.",
    "The database session remains active although all assigned ratings have been submitted; this fixed snapshot is dated.",
    "Per-role and per-category slices are exploratory and do not establish general routing rules.",
    "Seedream endpoint does not expose a seed; no deterministic regeneration claim is made.",
    "Prompt optimization was standard; the service may rewrite the supplied prompt. Exact anchor endpoint/settings are not recorded in this session.",
    "No direct comparisons with photography, Seedream 4.0, Qwen Image 2.0 Pro or GPT Image 2.5 Sunburst were run in this session."
  ]
}
