FLUX.2 [max] vs Wan 2.6
Head-to-head across 11 challenges
FLUX.2 [max]
52.4%
win rate
Ties
9.5%
Wan 2.6
38.1%
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
FLUX.2 [max]
- + Excellent preservation of the specific car model and details
- + Identifiable facial features and hairstyle matching the source
- + Accurate depiction of the California coastline environment
- − The man appears to be on the passenger side for a US-spec car, or the image is at an odd angle for driving
- − The man's expression is very stern compared to the source
Wan 2.6
- + Captures a much more dynamic sense of motion with blurred wheels and background
- + Maintains the man's cheerful expression from the source image
- + Stronger composition that emphasizes the 'driving' aspect
- − Notable distortion of the car's proportions (the front end is shortened and stylized)
- − The man's hand on the steering wheel has anatomical issues (extra long/unstructured fingers)
Verdict: FLUX.2 [max] did a superior job of preserving the physical characteristics of both the car and the man, maintaining the specific Rolls-Royce model accurately. However, Wan 2.6 provided a much more creative and visually pleasing composition by adding motion blur and keeping the subject's original smile, even though it heavily distorted the car's anatomy. FLUX.2 [max] is the winner for its technical accuracy in an editing task.
Modern Clean Menu
Text-to-Image“Modern minimalist restaurant menu design, white background with colorful food photos in grid, sections for appetizers/pizza/mains, bold sans-serif fonts, vibrant accents, clean professional layout for casual dining.”
AI Judge Analysis
FLUX.2 [max]
- + Excellent typographic hierarchy with clean, bold sans-serif fonts
- + Very organized grid layout and clear sectioning
- + Logical pricing alignment and professional icons
- − Internal logic error: includes pizza photos under the 'Appetizers' heading
- − The text is gibberish/placeholder language
Wan 2.6
- + Strong 'vibrant accents' with the colorful geometric borders
- + High-quality, appetizing photography in the grid
- + Includes clear dollar signs for pricing
- − Layout feels a bit cluttered with redundant 'Pizza' headings
- − Several text rendering errors and overlapping characters
- − Section headers are less prominent than in the other model
Verdict: FLUX.2 [max] creates a much more professional and realistic menu layout that feels like a finished graphic design product, despite its confusion over which food photos go in which section. Wan 2.6 has more vibrant color accents and excellent photography, but suffers from significantly more text artifacts and a slightly less intuitive information hierarchy.
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
FLUX.2 [max]
- + Excellent preservation of the original face and skin texture
- + Hair lighting matches the environmental light perfectly
- + Highly realistic stray hairs and flyaways
- − The hairline on the forehead looks slightly superimposed compared to the original skin
- − Smaller volume of hair compared to the request for 'full, thick'
Wan 2.6
- + Successfully added a very full, thick head of hair
- + Maintains the subject's overall likeness and facial structure
- + Matches hair color to the existing beard well
- − Subtle changes to the eyelid/eyebrow area compared to the source
- − The hair texture is slightly softer and more painterly than the sharp details of the source image
- − Small artifact where the hair meets the left temple
Verdict: FLUX.2 [max] did an incredible job of preserving the original image's fidelity, making the hair look naturally grown, though the volume is a bit conservative. Wan 2.6 provided a much thicker, more stylistic hairstyle that fits the 'full' requirement well, but it introduced very slight changes to the eyes and skin texture that drift away from the source image. FLUX.2 [max] is the winner for its superior realism and flawless preservation of the original's lighting and detail.
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
FLUX.2 [max]
- + Excellent adherence to the 'diorama base' prompt with a multi-tiered platform.
- + Very clean and professional typography with perfect spacing for 'JAPAN' and 'SUSHI'.
- + Soft, refined textures on the wooden board and sushi elements look high-quality.
- − The placement of the sushi rolls is slightly off-center on the plate.
- − The textures are a bit flat compared to the requested realistic PBR materials.
Wan 2.6
- + Better material rendering for the fish and rice, showcasing more 'PBR' style details.
- + Bolder, more vibrant colors that pop against the blue background.
- + Includes a shrimp (ebi) Nigiri which adds visual variety to the dish.
- − The text layout is slightly cluttered with the flag icon squeezed between the words.
- − The diorama base is a simple block, less interesting than the tiered design in Model A.
Verdict: Both models followed the prompt exceptionally well, producing clean, isometric 3D scenes. FLUX.2 [max] wins on composition and layout, particularly with its sophisticated tiered diorama base and cleaner typography, whereas Wan 2.6 has slightly better material textures for the food itself but a less balanced text arrangement.
Night Sky Transformation
Editing“Change the scene to night: a deep, dark sky with subtle, glistening stars visible behind the mountain.”
AI Judge Analysis
FLUX.2 [max]
- + Excellent source preservation, keeping the town layout nearly identical to the original.
- + Realistic lighting and color palette for a moonlit or deep twilight mountain scene.
- + Subtle, natural-looking stars as requested in the prompt.
- − The sky is a bit dark/flat in the upper corners.
- − Lighting on the mountain peak is slightly inconsistent with a total night sky.
Wan 2.6
- + Very clearly defined stars that strongly adhere to the 'glistening' request.
- + Dramatic night atmosphere with high contrast.
- − The stars appear somewhat artificial and uniform, like a pattern overlay.
- − Some texture loss in the mid-ground mountain slopes compared to the source.
Verdict: Both models followed the instructions well, but FLUX.2 [max] produced a much more realistic and cohesive image that better preserved the fine details of the original town. While Wan 2.6 provided more prominent stars, they felt somewhat pinned onto the background, whereas the stars and lighting in FLUX.2 [max] felt integrated into the scene.
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
FLUX.2 [max]
- + Excellent narrative composition combining all elements into a unified scene.
- + Strong preservation of facial features from the source image in the caricature style.
- + High level of detail in the background hockey rink and themed dog jerseys.
- − The denim jacket, while from the source, feels a bit casual for a 'TV anchor' persona compared to Model B.
Wan 2.6
- + Effective 'exaggerated' facial expression that fits the caricature request.
- + Clearly depicts both the TV studio setting and the anchor profession with a suit and headset.
- + Integrates the hockey stick and puck directly into the foreground action.
- − The facial resemblance to the source image is significantly weaker than Model A.
- − Structural issues with the hockey stick passing behind the character's arm but appearing in front of the hand.
Verdict: FLUX.2 [max] is the winner because it successfully carries over the subject's specific facial features into the caricature while brilliantly weaving the news anchor, hockey, and dogs into a single coherent 'Hockey Night' broadcast concept. Wan 2.6 captures the 'exaggerated' spirit well and uses more professional attire, but it loses the likeness of the original woman and has some confusing perspective issues with the hockey stick.
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
FLUX.2 [max]
- + Excellent Ghibli-inspired aesthetic with warm, grainy paper texture.
- + Preserves the composition and poses of the iconic meme perfectly.
- + Soft, pastel color palette and gentle lighting match the prompt beautifully.
- − The woman in the foreground's facial expression is a bit generic compared to the original.
Wan 2.6
- + High-quality watercolor texture with visible hand-painted linework.
- + Great character designs that balance realism with a Studio Ghibli anime style.
- + Strong adherence to the requested high-key lighting for a dreamy look.
- − Added white sparkles/dust can feel a bit distracting from the main subject.
- − The man's expression is slightly more neutral/sad than the 'distracted' look of the original.
Verdict: Both models did an exceptional job translating a famous photographic meme into a specific art style while maintaining the composition. FLUX.2 [max] captures the nostalgic, warm atmosphere of a 90s Ghibli film more accurately, whereas Wan 2.6 leans more toward a modern watercolor manga illustration style. FLUX.2 [max] is the winner for its superior blend of lighting and texture that feels more authentic to the requested 'warm, nostalgic mood'.
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
FLUX.2 [max]
- + Successfully creates a convincing Duchenne smile with natural teeth and eye crinkles.
- + High degree of source preservation for hair, lighting, and background.
- + Excellent skin texture and realistic transitions around the mouth.
- − Slightly changes the bridge and tip of the nose, making it appear a bit broader than the source.
Wan 2.6
- + Exceptional preservation of the original nose shape and facial structure.
- + Perfectly captures the eye crinkles and cheek raise requested.
- + Maintains the exact skin texture, freckles, and lighting of the source.
- − The alignment of the upper teeth is slightly asymmetric compared to the head tilt.
Verdict: Both models performed exceptionally well at a difficult image editing task. FLUX.2 [max] produces a very warm and convincing smile but slightly alters the nose shape. Wan 2.6 is the winner because it managed to add a complex expression while keeping the subject's unique facial features (especially the nose and eye shape) almost perfectly identical to the source image.
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
FLUX.2 [max]
- + Successfully added a large volume of flying leaves across the scene.
- + Clearly depicted hair blowing strongly in the wind.
- + Maintained the overall composition and character likeness well.
- − Introduced a slight mutation to the subject's left hand (fingers appear messy).
- − The leaves are static and lacks motion blur, making them look like they are 'stuck' to the air.
Wan 2.6
- + Excellent hair physics that feels more natural and voluminous.
- + Better preservation of the subject's anatomy, including the hands.
- + Added fewer but more tastefully placed flying leaves.
- − Fewer leaves than requested, which reduces the 'energetic and lively' feel compared to the other model.
Verdict: Both models followed the instructions well, successfully adding wind-blown hair and flying leaves while keeping the source image mostly intact. FLUX.2 [max] captures the 'lively' energy better with a high density of leaves, but it introduces a visible artifact on the woman's left hand. Wan 2.6 is the superior choice because it achieves the requested motion effect with much higher global coherence and preserved details, despite having fewer flying leaves.
Vintage Cafe Logo
Text-to-Image“Vintage minimalist restaurant logo for "Caffè Florian", retro cloche dome with steam and "Est. 1720" banner, classic typography, warm brown and cream tones, subtle texture on light background, vector emblem style.”
AI Judge Analysis
FLUX.2 [max]
- + Perfect text rendering for both the name and the date banner.
- + Excellent composition with a professional emblem circular frame.
- + Subtle, realistic paper texture that adds to the vintage feel.
- − The line weight of the steam is a bit thin compared to the rest of the illustration.
Wan 2.6
- + Strong contrast and bold vector style.
- + Accurate text rendering of the brand name.
- − The 'Est. 1720' banner is awkwardly placed and partially overlaps the cloche.
- − Heavy grunge texture on the edges feels a bit generic compared to the requested minimalist style.
- − Composition is less balanced than the circular emblem.
Verdict: FLUX.2 [max] produced a much more polished and professional logo that perfectly followed all prompt instructions, including the specific banner request. Wan 2.6 created a decent graphic but struggled with the placement of the banner and the overall balance of a minimalist emblem.
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
FLUX.2 [max]
- + Follows the complex 6-step infographic structure perfectly with clear icons for each stage.
- + Excellent text rendering with accurate spelling for 'Apollo 11 Mission' and astronaut names.
- + Consistently follows the flat-vector style and NASA-inspired color palette.
- − Minor spelling error in 'Tranquiity' (missing 'i') and 'Translunar' icon shows a moon instead of a trajectory arc.
- − The sequence of steps in the grid is slightly non-linear, jumping from top-right to middle-left.
Wan 2.6
- + Clean, minimalist aesthetic with a good color palette.
- + Correctly identifies the three astronauts by name.
- − Completely failed to include the requested 6-step infographic content.
- − The text 'ARMSTRONG' overlaps with the silhouette edges, reducing legibility.
- − The design is more of a book cover than an informative infographic poster.
Verdict: FLUX.2 [max] significantly outperformed Wan 2.6 by actually creating the requested 6-step infographic with high-quality icons and accurate labels. Wan 2.6 ignored almost all the specific instructions regarding the mission steps, resulting in a minimalist poster that lacks the required complexity and educational value.
FLUX.2 [max]
Black Forest Labs' flagship image generation model delivering state-of-the-art quality with exceptional realism, precision, and consistency for both text-to-image and advanced image editing
Wan 2.6
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