The ± value shows how much the counted judges’ scores vary across both rounds of scoring.
Claude Opus 5.5 · medium · Pass 17.10
Claude Opus 5.5 · medium · Pass 27.40
GPT-6 Astra · medium · Pass 17.20Not counted
GPT-6 Astra · medium · Pass 27.30Not counted
Grok 4.7 · medium · Pass 16.20
Grok 4.7 · medium · Pass 25.85
Scoring criteria
Pelican likeness
8.25
Bicycle correctness
7.50
Riding pose
5.25
Composition
6.25
Obvious errors (higher is better)
4.75
Drawing checks passedNormal
Judges
Claude Opus 5.5 · medium · Pass 1
7.10
What the judge sees: Original (Chinese) 这是一幅白色背景上的扁平风格插画，画的是一只白色鹈鹕骑在自行车上。鹈鹕的身体是椭圆形，带一只浅灰色翅膀，脖子细长并向上弯曲，头朝右，长着一张大大的橙色喙和喉囊；它伸出灰色翅膀尖握住车把，两条细长的橙色腿向下伸到脚踏处。自行车是青绿色三角形车架，车把是深蓝色，前后两个深蓝色车轮上有浅灰色辐条，车子朝向画面右侧。
The pelican looks most like a real bird, but the legs are crossed and twisted, there's no saddle, and there's too much empty space at the top.
Machine translated
Original (Chinese) 鹈鹕造型最像真鸟，但双腿交叉扭曲、没有车座，画面上方留白过多。
Long beak and throat pouch: Yes
Two wheels: Yes
Frame connects the wheels: Yes
Bird sits on the seat: No
Feet on the pedals: Yes
Pelican likeness
9.00
Bicycle correctness
8.00
Riding pose
6.00
Composition
6.00
Obvious errors (higher is better)
5.00
Claude Opus 5.5 · medium · Pass 2
7.40
What the judge sees: Original (Chinese) 这是一幅白色背景上的扁平风格插画，画的是一只白色鹈鹕骑在自行车上。鹈鹕的身体是椭圆形，带一只浅灰色翅膀，脖子细长并向上弯曲，头朝右，长着一张大大的橙色喙和喉囊；它伸出灰色翅膀尖握住车把，两条细长的橙色腿向下伸到脚踏处。自行车是青绿色三角形车架，车把是深蓝色，前后两个深蓝色车轮上有浅灰色辐条，车子朝向画面右侧。
The pelican and bicycle build both look good, but the legs are crossed and twisted, with too much empty space at the top.
Machine translated
Original (Chinese) 鹈鹕造型和自行车结构都不错，但双腿交叉扭曲，上方留白过多。
Long beak and throat pouch: Yes
Two wheels: Yes
Frame connects the wheels: Yes
Bird sits on the seat: Yes
Feet on the pedals: Yes
Pelican likeness
8.00
Bicycle correctness
8.00
Riding pose
8.00
Composition
6.00
Obvious errors (higher is better)
6.00
GPT-6 Astra · medium · Pass 1
7.20Not counted
What the judge sees: Original (Chinese) 白色背景上，一只浅米色的鸟位于一辆自行车上方，鸟头朝右，长喙呈橙黄色，翅膀为浅灰色，两条橙色的腿伸向车架下部。自行车的车架呈青绿色，左右两个圆形车轮有深蓝灰色轮胎和浅灰色辐条，右上方有一段深蓝灰色的横向车把。
Original (Chinese) 握把与踩踏关系较明确，但缺少可辨认的车座。
Long beak and throat pouch: Yes
Two wheels: Yes
Frame connects the wheels: Yes
Bird sits on the seat: No
Feet on the pedals: Yes
Pelican likeness
8.00
Bicycle correctness
7.00
Riding pose
6.00
Composition
8.00
Obvious errors (higher is better)
7.00
GPT-6 Astra · medium · Pass 2
7.30Not counted
What the judge sees: Original (Chinese) 白色背景上，一只浅米色的鸟位于一辆自行车上方，鸟头朝右，长喙呈橙黄色，翅膀为浅灰色，两条橙色的腿伸向车架下部。自行车的车架呈青绿色，左右两个圆形车轮有深蓝灰色轮胎和浅灰色辐条，右上方有一段深蓝灰色的横向车把。
Original (Chinese) 鹈鹕特征清楚，脚与踏板相接，但缺少明确车座。
Long beak and throat pouch: Yes
Two wheels: Yes
Frame connects the wheels: Yes
Bird sits on the seat: No
Feet on the pedals: Yes
Pelican likeness
9.00
Bicycle correctness
7.00
Riding pose
6.00
Composition
7.00
Obvious errors (higher is better)
7.00
Grok 4.7 · medium · Pass 1
6.20
What the judge sees: Original (Chinese) 白色背景上有一只浅米色鹈鹕骑在一辆自行车上。鹈鹕身体圆润，翅膀浅灰，长颈向上弯曲，橙色大嘴朝右张开，橙色细腿交叉踩在踏板上。自行车车架为青绿色，两轮为深色轮廓并带浅色辐条，车把位于鹈鹕前方。
Pelican and frame are clear, but the legs cross and the feet aren't on the pedals.
Machine translated
Original (Chinese) 鹈鹕与车架清楚，但双腿交叉，脚没有踩踏板。
Long beak and throat pouch: Yes
Two wheels: Yes
Frame connects the wheels: Yes
Bird sits on the seat: Yes
Feet on the pedals: No
Pelican likeness
8.00
Bicycle correctness
7.00
Riding pose
4.00
Composition
7.00
Obvious errors (higher is better)
4.00
Grok 4.7 · medium · Pass 2
5.85
What the judge sees: Original (Chinese) 白色背景上有一只浅米色鹈鹕骑在一辆自行车上。鹈鹕身体圆润，翅膀浅灰，长颈向上弯曲，橙色大嘴朝右张开，橙色细腿交叉踩在踏板上。自行车车架为青绿色，两轮为深色轮廓并带浅色辐条，车把位于鹈鹕前方。
Long neck with a pouch, but the legs cross in midair and the bird isn't sitting on the saddle.
