Reve AI's text-to-image generation model with strong aesthetic quality, accurate text rendering, and detailed instruction following capabilities
Settled by community votes across 8 shared challenges, with an AI judge weighing in on each.
Reve Image 1.0
#42 of 62 in Text-to-Image
Wan 2.6
#28 of 62 in Text-to-Image
Where the votes landed
Reve Image 1.0
46.2%
win rate
Ties
0.0%
Wan 2.6
53.8%
win rate
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
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
Reve Image 1.0
- + Excellent photorealistic lighting and shallow depth of field.
- + Sleek, modern minimalist aesthetic.
- + High resolution with clean, sharp edges on the glass cube.
- − The plant is to the side rather than behind the cube, failing the positioning prompt.
- − The blue sphere appears to be floating mid-air inside the cube without support.
- − The sphere is not clearly visible 'through' the glass relative to the plant background.
Wan 2.6
- + Perfect adherence to spatial instructions with the plant behind the cube.
- + Realistic texture on the weathered red book and wooden table.
- + Accurate physics with the sphere resting on the bottom of the cube.
- − The glass cube has some structural inconsistencies in its internal reflections.
- − Lighting on the plant feels slightly disconnected from the window light source.
Verdict: Wan 2.6 is the superior image because it correctly places the green plant behind the cube as requested, whereas Reve Image 1.0 places it to the left. Additionally, Wan 2.6 provides a more grounded sense of physics and realistic textures, specifically in the weathered detail of the red book and the sphere's placement.
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
Reve Image 1.0
- + Excellent engraving detail on the armor piece.
- + Captures the warm torchlight reflection very naturally across the face and metal.
- − The character looks significantly younger than a typical 'battle-worn' paladin.
- − The 'beads' in the hair look more like metal clips or staples.
Wan 2.6
- + Perfectly captures the 'battle-worn' aesthetic with realistic dirt, grime, and scars.
- + Superior texture on leather straps, cloth underlayers, and chainmail.
- + The hair braiding with beads is executed more artistically and accurately to the prompt.
- − The bokeh sparks in the background are a bit large and slightly distracting.
Verdict: While Reve Image 1.0 produces a clean and high-quality image, it fails to capture the 'battle-worn' essence of the prompt, featuring a character who looks like a teenager in clean armor. Wan 2.6 provides much better adherence to all prompt details, specifically the specialized textures of the leather and cloth, the scars, and the weathered appearance of the equipment.
Bald man challenge
Image 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
Reve Image 1.0
- + Perfectly preserves facial features and fine skin details
- + Excellent lighting match between the new hair and the existing scene
- + Highly accurate preservation of the original image background and clothing
- − The hair texture looks slightly frizzy and overly voluminous for the style
- − The hairline integration with the forehead is a bit harsh
Wan 2.6
- + Realistic, sleek hair texture and styling
- + Better integration of hair around the ears
- + Succesfully preserves identity and image background
- − Slightly alters the eyebrows, making them bushier than the source
- − The glasses frames are slightly warped compared to the original
Verdict: Both models performed excellently at this image editing task, preserving the person's identity and the surrounding environment almost perfectly. Reve Image 1.0 is the winner because it strictly preserved the original facial features (like the eyebrows and glasses geometry) better than Wan 2.6, even though Wan 2.6 provided a more modern and well-groomed hairstyle.
Isometric Miniature Diorama Scenes
Text-to-Image“Create a clear, 45° top-down isometric miniature 3D cartoon scene of Japan's signature dish: sushi, with soft refined textures, realistic PBR materials, gentle lighting, on a small raised diorama base with minimal garnish and plate. Solid light blue background. At top-center: 'JAPAN' in large bold text, 'SUSHI' below it, small flag icon. Perfectly centered, ultra-clean, high-clarity, square format.”
AI Judge Analysis
Reve Image 1.0
- + Excellent soft, refined textures on the salmon
- + The rounded, polished 3D aesthetic fits the 'cartoon miniature' request perfectly
- + Clean and legible text layout with a accurate flag icon
- − The tuna piece glows a bit too much, losing the PBR material feel
- − Layout feels slightly less 'isometric' than Model B
Wan 2.6
- + Perfect 45-degree isometric composition with a clear diorama base
- + Excellent variety of sushi items including ebi (shrimp) and garnishes
- + Highly effective use of PBR materials with realistic wood grain and soft shadows
- − The flag icon is placed to the left of 'SUSHI' rather than being a small icon as part of a top-center cluster
- − The rice texture is slightly repetitive and chunky
Verdict: Reve Image 1.0 produces a very cute, toy-like aesthetic with superior text rendering, but Wan 2.6 better captures the specific 'isometric diorama' requirement of the prompt. Wan 2.6 feels like a more complete miniature scene with a professional 3D render feel, despite the slight deviation in text placement.
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
Reve Image 1.0
- + Excellent preservation of the subject's facial features and likeness.
- + Clever integration of hobbies, such as the hockey jersey under the blazer and the dog wearing a headset.
- + Vibrant colors and a clean, digital caricature style.
- − The transition between the neck and the blazer looks a bit disconnected.
- − The hockey stick is floating/awkwardly placed on the desk.
Wan 2.6
- + Strong 'chibi' or cartoon style that is very cohesive.
- + Good composition with a clear background set.
- + Includes all requested elements clearly.
- − Poor preservation of the original person's likeness; the face looks generic.
- − The hand holding the microphone is poorly rendered with too few fingers.
Verdict: Reve Image 1.0 is the winner because it successfully creates a caricature while maintaining the recognizable facial features of the woman in the source image. Wan 2.6 provides a fun cartoon, but the character no longer resembles the original person, failing a key aspect of a personalized caricature edit.
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
Reve Image 1.0
- + Perfectly captures the Studio Ghibli cel-shaded animation style.
- + Preserves the composition and poses of the iconic meme while translating them to anime.
- + The background redesign feels very much like a Ghibli village scene.
- − The foreground character's face is very blurry, mimicking the original depth of field too literally for an illustration.
Wan 2.6
- + Beautiful watercolor/pastel aesthetic with high detail.
- + Successfully maintains the source image composition and the specific facial expressions of the subjects.
- + The soft lighting and 'shoujo' sparkles create a very dreamy mood.
- − The style leans more towards modern manga watercolor illustration rather than the specific Studio Ghibli hand-painted film aesthetic requested.
- − Less stylized than Model A, remaining closer to the original proportions of the photo.
Verdict: Both models did an excellent job of preserving the composition of the 'distracted boyfriend' meme while applying a new style. Reve Image 1.0 is the winner for prompt adherence, as it perfectly mimics the line work, character designs, and background painting style synonymous with Studio Ghibli films. While Wan 2.6 is a beautiful illustration, its style is more generalized watercolor manga and doesn't capture the specific Ghibli 'soul' as accurately as Reve Image 1.0.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
Reve Image 1.0
- + Successfully adds dynamic motion to the hair with realistic flow.
- + Adds a large quantity of leaves that create a strong sense of wind.
- + Maintains the subject's identity and facial features perfectly.
- − The sheer number of leaves is slightly overwhelming and covers parts of the subjects.
- − Some leaves have a motion blur that looks a bit digital/artificial.
Wan 2.6
- + Achieves a very natural wind-blown effect on the hair.
- + Subtle and tasteful addition of leaves that doesn't distract from the subjects.
- + High level of preservation of the original image details and lighting.
- − The 'energetic and lively' feel is a bit more muted compared to the prompt's request.
- − Fewer leaves make the motion feel less 'dynamic' than Model A.
Verdict: Reve Image 1.0 followed the prompt more aggressively, providing a high-energy scene with significant hair movement and many flying leaves, though the quantity of leaves feels a bit cluttered. Wan 2.6 provided a cleaner, more realistic edit with beautiful hair movement, but it feels less 'dynamic' overall. Reve Image 1.0 is the winner for better capturing the intended mood and specific instructions of the edit.
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
Reve Image 1.0
- + Successfully included multiple infographic steps as requested.
- + Handled the NASA-inspired color palette and flat-vector style perfectly.
- + The iconography is consistent and relevant to the mission steps.
- − Several spelling errors in the text (e.g., 'Anmlot', 'Lunnr', 'Dessent').
- − Missing the 'Translunar' step requested in the prompt.
Wan 2.6
- + Clean typography and good use of white space.
- + Creative use of astronaut profiles to label the crew.
- − Failed to include the requested infographic steps/process entirely.
- − The graphic is a cover or poster rather than the requested infographic.
Verdict: Reve Image 1.0 adhered much better to the complex prompt's structural requirements, providing a sequence of icons and labels for the mission steps, even though it suffered from typological errors. Wan 2.6 created a visually pleasing poster but completely ignored the core request for a multi-step infographic. Reve Image 1.0 is the winner for following the instructional intent.
Explore each model
Alibaba's multimodal generation model from the Wan AI suite, supporting text-to-video, image-to-video, reference-to-video with audio, and text-to-image, in both Chinese and English