GPT Image 1.5 vs Reve Image 1.0
Head-to-head across 10 challenges
GPT Image 1.5
75.0%
win rate
Ties
8.3%
Reve Image 1.0
16.7%
win rate
Challenge Results
Geometric Composition
Text-to-Image“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”
AI Judge Analysis
GPT Image 1.5
- + Excellent adherence to the glass cube geometry and material properties.
- + Very high detail on the red book's texture and the blue sphere's reflection.
- + Succesfully captures the plant being visible 'through' the glass as requested.
- − The blue sphere is quite large relative to the cube, rather than 'small' as prompted.
Reve Image 1.0
- + Accurately renders a 'small' blue sphere relative to the cube size.
- + Good composition with a clear soft window light effect casting a shadow on the table.
- − The blue sphere appears to be floating mid-air inside the cube without support.
- − The perspective of the cube is slightly warped, and the glass lacks the thickness and caustic realism seen in Model A.
Verdict: GPT Image 1.5 is the superior image due to its exceptional material realism and detail, particularly in how the glass cube interacts with its environment and the red book. While Reve Image 1.0 followed the size constraint for the sphere better, the sphere's floating state and the less convincing glass rendering make it the weaker choice.
Fantasy Warrior
Text-to-Image“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”
AI Judge Analysis
GPT Image 1.5
- + Exceptional textural detail on the skin, leather straps, and engraved armor.
- + Strong adherence to the 'battle-worn' aesthetic with realistic dirt and scars.
- + Powerful composition with high-quality bokeh and cinematic lighting.
- − The hair beads are somewhat merged with the hair strands in a less distinct way.
Reve Image 1.0
- + Excellent representation of the beaded braids requested in the prompt.
- + Ornate engraving on the plate armor is clearly visible and well-executed.
- + Good use of warm torchlight and shallow depth of field.
- − The character looks very young and lacks the 'battle-worn' texture requested.
- − The skin texture is too smooth and the dirt looks like individual isolated dots rather than realistic grime.
Verdict: GPT Image 1.5 is the superior image due to its incredible attention to detail, specifically the lifelike eyes and the realistic 'battle-worn' texture of the skin and armor which perfectly captures the prompt's mood. While Reve Image 1.0 followed the specific request for hair beads more literally, it failed to deliver on the gritty, weathered aesthetic required for a battle-worn paladin.
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
- + Natural integration of the hair with the existing sideburns and beard
- + High level of preservation for facial features and clothing
- + Realistic texture and density that matches the lighting of the scene
- − The hair slightly changes the upper forehead shape and skin texture
Reve Image 1.0
- + Excellent preservation of the background and original clothing
- + Follows the prompt for a 'full' head of hair
- − The hair volume appears unnaturally high and lacks a realistic silhouette
- − The blending at the temples and sideburns is messy and less coherent than the other model
- − The image overall appears slightly softened/less sharp than the source
Verdict: GPT Image 1.5 provides a much more convincing and realistic edit, blending the new hair seamlessly into the subject's existing beard and sideburns while maintaining the photo's original sharp quality. Reve Image 1.0 creates an unnaturally large, 'wig-like' volume of hair that doesn't integrate well with the subject's facial structure or the surrounding environment.
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
- + Perfectly captures the 'deep dark sky' and 'subtle glistening stars' prompt.
- + Effectively updates the lighting on the slopes and valley to match the new night sky.
- + Maintains the structural integrity of the mountains and village while shifting the mood.
- − The sky feels slightly separate from the mountain edge in a few spots (minor masking artifacts).
Reve Image 1.0
- + Preserves the lighting of the village lights and river reflections very well.
- + Highly accurate preservation of the foreground and mid-ground mountain geometry.
- − The stars are extremely sparse and barely visible, making the sky look empty.
- − The transition between the mountain peak and the dark sky is a bit harsh.
Verdict: GPT Image 1.5 followed the prompt much more effectively, providing a beautiful, star-filled night sky that feels integrated with the landscape. Reve Image 1.0 preserved more of the original image's details but failed to deliver the 'glistening stars' requested in the prompt, resulting in an unnaturally black sky.
Over-the-top cartoon caricature
Editing“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
AI Judge Analysis
GPT Image 1.5
- + Excellent caricature style with exaggerated features that still look like the person.
- + Superb prompt adherence by including multiple dogs, hockey players/gear, and a detailed news set.
- + Very clean text rendering and high-quality artistic detail.
- − The small dog's helmet and stick are slightly messy in terms of fine detail.
Reve Image 1.0
- + Strong resemblance to the original person's facial features.
- + Creative elements like the dog wearing headphones and a tie.
- + Good clarity and clean composition.
- − The hockey stick is just floating/propped up awkwardly without much context.
- − Less 'exaggerated' as a caricature compared to Model A.
Verdict: GPT Image 1.5 followed the prompt much more effectively, creating a busy, humorous, and highly detailed caricature that perfectly blended the news anchor role with dogs and hockey themes. Reve Image 1.0 felt a bit more like a standard illustration with fewer elements, and the facial proportions were less 'exaggerated' in the traditional caricature sense.
Victorian Greenhouse Oasis
Text-to-Image“Hyper-photorealistic interior of a lush Victorian glass greenhouse filled with exotic tropical plants, vibrant blooming orchids, tall ferns, colorful butterflies in flight, sunlight filtering through ornate glass roof creating realistic caustics and dew on leaves, intricate iron framework visible, misty atmosphere, 8K masterpiece.”
AI Judge Analysis
GPT Image 1.5
- + Expertly captured the misty atmosphere and god rays requested in the prompt.
- + Highly detailed Victorian ironwork that feels authentic to the architectural style.
- + Exceptional rendering of dew drops on leaves and realistic lighting caustics.
- − Butterflies are slightly sharp compared to the atmospheric depth of the background.
- − The composition is very busy, though it fits the 'lush' description.
Reve Image 1.0
- + Vibrant colors and clear representation of various tropical ferns and orchids.
- + Good use of a symmetrical central composition under the glass dome.
- − Lacks the atmospheric mist and realistic light filtration requested in the prompt.
- − Several butterflies appear distorted or have unnatural shapes and placements.
- − The lighting feels flat and digital rather than 'hyper-photorealistic'.
Verdict: GPT Image 1.5 is the clear winner as it successfully translated complex lighting prompts like 'misty atmosphere' and 'sunlight filtering' into a cohesive, masterpiece-quality image. In contrast, Reve Image 1.0 failed to generate the requested atmosphere and ironwork details, resulting in a flatter image with several anatomical errors in the butterflies.
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
- + Excellent preservation of the source image's composition and poses.
- + Beautiful, warm color palette that fits the 'nostalgic' and 'dreamy' prompt keywords.
- + Maintains the distinct facial expressions of the original meme perfectly in a new style.
- − The art style leans more towards modern shoujo anime or digital watercolor than the specific hand-drawn Ghibli aesthetic.
- − Overuse of soft glow/haze can obscure some of the fine linework.
Reve Image 1.0
- + Very accurate capture of the Studio Ghibli '90s cel-shaded look and line weights.
- + Excellent hand-painted background textures that feel grounded in the requested style.
- + Effective use of the 'out of focus' effect for the foreground character to match the source photo's depth of field.
- − The man's facial expression is a bit more neutral/blank compared to the source's 'pucker' expression.
- − The woman in red's face is a bit generic and lacks the personality of the source.
Verdict: Both models did an excellent job translating the 'Distracted Boyfriend' meme into an illustration. GPT Image 1.5 followed the mood and lighting instructions perfectly, creating a beautiful image, but Reve Image 1.0 was much more successful at specifically capturing the Studio Ghibli art style and 'painted' background aesthetic requested.
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 adherence to the Duchenne smile request, including natural teeth and eye crinkles.
- + Near-perfect preservation of the background, hair, and clothing.
- + Realistic skin folding and cheek shadows associated with a genuine smile.
- − Slightly alters the iris color and brightness, making the eyes appear a bit greener/lighter than the source.
Reve Image 1.0
- + Perfectly preserves almost every pixel of the original image outside the immediate mouth/cheek area.
- + Maintains the exact eye color and shape from the source image.
- − Failed to include the requested 'natural teeth' in the smile.
- − The smile looks somewhat forced or 'smirking' rather than a genuine Duchenne smile.
Verdict: GPT Image 1.5 followed the complex instructions for a Duchenne smile much more effectively, providing the requested teeth and natural facial crinkles, though it slightly brightened the iris color. Reve Image 1.0 was superior at preserving the exact original pixels and eye color, but the resulting smile was closed-mouthed and less 'genuine' than requested.
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 person's face and clothing.
- + Subtle and realistic outward-blowing hair effect.
- + The leaves are integrated into the scene's lighting well.
- − The leaf density is a bit high, which can feel slightly cluttered.
- − The leash handle area has some minor artifacting where it meets the hand.
Reve Image 1.0
- + Strong sense of dynamic motion with larger, blurred foreground leaves.
- + Very energetic hair-blowing effect that matches the wind direction of the leaves.
- + Good preservation of the dog and background elements.
- − The larger leaves in the foreground are a bit blurry and distract from the subject.
- − The hair edit looks slightly more 'painted on' compared to the original texture.
Verdict: Both models did an excellent job of following the edit instructions while preserving the source image. GPT Image 1.5 is preferred for its superior preservation of the subjects' fine details and more natural integration of the leaves, whereas Reve Image 1.0 has slightly more aggressive motion blur that obscures the foreground too much.
Apollo 11: Journey to Tranquility
Text-to-Image“Create a clean, modern vector infographic poster about the Apollo 11 mission. NASA-inspired palette (navy, white, muted red, light gray). Flat-vector style, crisp lines, consistent iconography, subtle gradients only. Steps (stop at landing): 1. Launch (Saturn Vicon) 2. Earth Orbit (Earth + orbit ring icon) 3. Translunar (trajectory arc icon) 4. Lunar Orbit (Moon + orbit ring icon) 5. Descent (lunar module descending icon) 6. Landing (lunar module on the surface icon) Small supporting elements (minimal text): • Crew strip: three silhouette icons with only last names: Armstrong, Aldrin, Collins. • Landing site marker: Moon pin labeled "Tranquility" only. Layout constraints: generous margins, large readable labels, clean background with subtle stars. Vector-only, print-poster look, high resolution.”
AI Judge Analysis
GPT Image 1.5
- + Excellent text rendering with correct spelling for all steps and names.
- + All six requested stages are clearly represented with logical, high-quality icons.
- + Consistent flat-vector style with a professional NASA-inspired palette.
- − The layout is a bit cramped at the top, cutting off the main title.
- − The 'Launch' section includes a globe icon that might be better suited for orbit.
Reve Image 1.0
- + Clean, minimalist layout with good use of negative space.
- + Accurately captures the requested flat-vector aesthetic.
- − Several spelling errors in icons (LUNNR, DESSENT, APQUO).
- − Missing the 'Translunar' step entirely, skipping from step 2 to 4.
- − Numbering is inconsistent and non-sequential.
Verdict: GPT Image 1.5 followed the prompt instructions perfectly, including all six specific mission steps with correct spelling and highly detailed vector icons. Reve Image 1.0 struggled with text accuracy and omitted the requested 'Translunar' stage, resulting in an incomplete and confusing infographic.
GPT Image 1.5
OpenAI's state-of-the-art image generation model with better instruction following and adherence to prompts
Reve Image 1.0
Reve AI's text-to-image generation model with strong aesthetic quality, accurate text rendering, and detailed instruction following capabilities