GPT Image 1.5 vs Qwen Image Edit 2511
Head-to-head across 6 challenges
GPT Image 1.5
67.6%
win rate
Ties
2.8%
Qwen Image Edit 2511
29.6%
win rate
Challenge Results
Man and Car in California
Editing“Make a photo of the man driving the car down the California coastline”
AI Judge Analysis
GPT Image 1.5
- + Excellent preservation of the specific Rolls-Royce car details and white exterior.
- + Seamlessly integrates the man and the car into the California coastline setting.
- + The man's appearance is perfectly consistent with the source image.
- − The perspective makes the man look slightly too small relative to the car seat/steering wheel.
Qwen Image Edit 2511
- + Successfully combines all elements from both source images.
- + The lighting on the man's face and clothing is very natural for an outdoor scene.
- + High clarity and vibrant colors in the coastline background.
- − Modified the car's interior with unnecessary wood paneling not found in the source.
- − The man's plaid coat is partially merged with his arm/torso in a slightly disjointed way.
Verdict: Both models did an excellent job of merging the two source images into a cohesive scene. GPT Image 1.5 is the winner because it better preserved the specific details of the source car, whereas Qwen Image Edit 2511 altered the interior dashboard significantly. GPT Image 1.5 also managed a more realistic integration of the man sitting 'inside' the vehicle's cabin.
Bald man challenge
Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
GPT Image 1.5
- + Excellent source preservation with no detectable changes to facial features.
- + The hair texture matches the existing beard perfectly.
- + The hairline is extremely realistic and integrated well with the forehead.
- − None notable.
Qwen Image Edit 2511
- + Successfully applied the requested edit of a full head of hair.
- + Preserved the background and clothing accurately.
- − The hair texture looks artificial and overly smooth compared to the beard.
- − Visible rendering artifacts where the hair meets the forehead/temples.
- − The hair volume and style look like an overlay rather than natural growth.
Verdict: GPT Image 1.5 is the clear winner as it seamlessly integrates the new hair with the existing person, matching the texture of the beard and maintaining a perfect natural hairline. Qwen Image Edit 2511 produces hair that looks like a digital overlay with a plastic-like texture that clashes with the realistic aesthetic of the original image.
Night Sky Transformation
Editing“Change the scene to night: a deep, dark sky with subtle, glistening stars visible behind the mountain.”
AI Judge Analysis
GPT Image 1.5
- + Excellent atmospheric lighting and realistic night sky.
- + Preserves the shape of the mountain and town layout perfectly.
- + High visual quality with realistic star distribution and Milky Way details.
- − The overall image is significantly darker than the source, losing some foreground texture.
Qwen Image Edit 2511
- + Successfully keeps the brightness of the village lights from the source image.
- + Accurately shifts the sky color to a deep blue.
- − The stars look like an artificial, repetitive grid of white dots.
- − The mountain peak loses its sharp detail compared to the source.
Verdict: GPT Image 1.5 provides a much more convincing and artistic night scene, rendering a realistic starry sky that integrates well with the landscape. Qwen Image Edit 2511 preserves more of the source image's light in the valley, but the star effect is very poor and looks like a basic overlay patterns rather than a night sky.
Studio Ghibli Anime Style
Editing“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
AI Judge Analysis
GPT Image 1.5
- + Successfully captures the Studio Ghibli art style with expressive faces and soft eyes
- + Excellent application of warm, nostalgic lighting and hand-painted texture
- + Strong adherence to the requested 'dreamy' atmosphere
- − Deforms some anatomical details, such as the man's hand and the woman's hair flow
- − Slightly less preservation of the original facial structures
Qwen Image Edit 2511
- + Excellent source preservation, keeping the characters' facial features very recognizable
- + Clean linework while maintaining the exact composition of the original photo
- − Fails to fully capture the Studio Ghibli aesthetic, appearing more like a standard digital tracing
- − The lighting and background lack the 'dreamy' and 'nostalgic' quality requested in the prompt
Verdict: GPT Image 1.5 is the winner because it successfully transforms the image into a completely different artistic style (Studio Ghibli) as requested, whereas Qwen Image Edit 2511 follows a more literal 'cartoon filter' approach. While Qwen preserves the original faces better, GPT Image 1.5 captures the specific mood, textures, and lighting of the requested animation style.
Neutral Expression to Genuine Smile
Editing{
"action": "image_edit",
"reference": "uploaded neutral portrait",
"change": "Warm genuine Duchenne smile: lips curved up, slight natural teeth, soft eye crinkles, subtle cheek raise",
"details": "Realistic smiling skin (dimples if present, soft cheek shadows), slightly brighter eyes; keep exact eye shape/color/iris",
"preserve_exact": "Face identity/structure, eyes/nose/lips/eyebrows, hair, skin texture/pores/freckles, makeup, clothing, head pose, background, lighting, shadows, framing",
"no_changes": "No face shape change, no new features, no gaze shift, no hair/clothing/lighting/background edits",
"style": "Ultra-photorealistic 8K portrait, sharp face focus, natural soft lighting, realistic skin glow"
}
AI Judge Analysis
GPT Image 1.5
- + Excellent preservation of the original image's texture and hair details.
- + Successfully implements a natural, high-quality smile that fits the face structure.
- − The eyes remain somewhat wide and neutral, lacking the requested Duchenne 'eye crinkles'.
- − Minimal change to the upper half of the face, making the smile feel slightly less genuine.
Qwen Image Edit 2511
- + Perfectly captures the 'Duchenne' smile with realistic eye crinkling and raised cheeks.
- + Maintains the skin texture and freckles while modifying the muscle movement correctly.
- + Brighter eyes as requested without changing the iris color.
- − Very minor softening of the skin texture on the cheeks compared to the original.
Verdict: GPT Image 1.5 does a great job of preserving the source image perfectly while adding a smile, but it misses the 'Duchenne' aspect of the prompt, leaving the eyes too neutral. Qwen Image Edit 2511 captures the genuine warmth requested, accurately rendering the eye crinkles and cheek raises while keeping the subjects identity intact.
Golden Hour Stroll
Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
GPT Image 1.5
- + Excellent preservation of the original face and clothing details.
- + Large number of flying leaves creates a strong sense of atmospheric motion.
- + Natural-looking hair flow that blends well with the existing lighting.
- − The leash handle has become slightly distorted compared to the source.
Qwen Image Edit 2511
- + Effective 'hair blowing' effect with more dramatic spread.
- + Very high fidelity in preserving the woman's facial features and the dog's expression.
- + Good use of scattered leaves to imply wind.
- − The hair strands are a bit more stylized and separated than the original texture.
- − Fewer leaves than Model A, making the environment feel slightly less 'energetic'.
Verdict: Both models performed excellently, maintaining near-perfect consistency with the source image's identity and background. GPT Image 1.5 is the winner because it added a greater volume of leaves which better fulfills the 'energetic and lively' part of the prompt, whereas Qwen Image Edit 2511 focused more heavily on the hair with a sparser distribution of leaves.
GPT Image 1.5
OpenAI's state-of-the-art image generation model with better instruction following and adherence to prompts
Qwen Image Edit 2511
Alibaba's Qwen image editing model for instruction-based image modifications and transformations