FLUX.2 [dev] Turbo vs GPT Image 1.5
Head-to-head across 10 challenges
FLUX.2 [dev] Turbo
44.4%
win rate
Ties
22.2%
GPT Image 1.5
33.3%
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
FLUX.2 [dev] Turbo
- + Excellent adherence to the 'small' sphere description, maintaining a realistic scale within the cube.
- + Superior lighting and atmosphere with a high level of photorealism in the wood grain and glass texture.
- + Higher visual depth and more natural integration of the plant behind the glass.
- − The plant is slightly more on the right than directly behind, though it is visible through the glass.
GPT Image 1.5
- + Perfectly follows the spatial instruction of placing the plant directly behind the cube.
- + Clean composition with very clear visibility of all requested elements.
- − The blue sphere is large rather than 'small' as requested in the prompt.
- − The lighting on the sphere feels a bit flat and CG-like compared to the rest of the image.
Verdict: FLUX.2 [dev] Turbo produced a more photorealistic image with better attention to the scale of the 'small' sphere. While GPT Image 1.5 placed the plant more accurately behind the cube, its sphere was too large and the overall lighting was less cinematic than FLUX.2.
Candid Street Photography
Text-to-Image“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent adherence to the 'motion blur' requirement for background cars.
- + Highly realistic skin texture and age-appropriate details for the subject.
- + Perfectly captures the 'reflections on wet pavement' and 'light rain' atmosphere.
- − The bicycle structure is physically incoherent, especially the front wheel and handlebars.
- − The subject's hands are mangled and blending into the bicycle frame.
GPT Image 1.5
- + Excellent 'shallow depth of field' and 'imperfect framing' which creates a candid feel.
- + Much more coherent bicycle mechanics and subject interaction.
- + Superior composition that feels more like a 50mm candid street shot.
- − Missed the 'motion blur from passing cars' request as the background car is sharp.
- − Slightly less emphasis on the 'light rain' visible falling through the air compared to Model A.
Verdict: FLUX.2 [dev] Turbo followed the movement and weather prompts more closely but failed significantly on mechanical and anatomical coherence. GPT Image 1.5 produced a much more believable and cinematic scene with better framing and realistic interaction, despite missing the motion blur requirement. GPT Image 1.5 is the winner for its superior visual quality and realism in the face of complex subject interaction.
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
FLUX.2 [dev] Turbo
- + Excellent depiction of colored beads in the braids.
- + The texture on the leather straps and the engraved metal is incredibly crisp and realistic.
- + Superior depth of field and bokeh spark integration.
- − The torch in the background is a bit blurry and slightly distracting.
- − The skin texture is a little smoother than requested for a 'battle-worn' subject.
GPT Image 1.5
- + Outstanding warm torchlight reflection across the metal and skin.
- + Stronger 'battle-worn' aesthetic with realistic sweat, dirt, and skin texture.
- + Ornate engraving on the armor looks very high-quality.
- − The 'braids with small beads' are interpreted more as metal rings/sleeves than beads.
- − The composition feels slightly more crowded compared to Model A.
Verdict: Both models performed exceptionally well, but FLUX.2 [dev] Turbo followed the specific detail of 'small beads' more accurately. GPT Image 1.5 offered superior lighting and a more convincing 'battle-worn' texture, but Model A's overall clarity and faithful reproduction of every prompt element make it the winner.
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 [dev] Turbo
- + Excellent graphic design elements with color-blocked accents.
- + Clean, modern grid layout for photos.
- + Professional typography for the main headers.
- − Nonsense filler text for menu items and descriptions.
- − Pricing is unrealistic (e.g., $260 for pizza).
- − Confusing photo placement where pizzas are shown under the 'Appetizers' heading.
GPT Image 1.5
- + Highly legible, accurate English text for menu items and descriptions.
- + Logical organization with photos corresponding to the categories.
- + Clean, professional white space usage.
- − Layout is slightly more traditional and less 'modern graphic' than Model A.
- − Section lines intersect photos at the bottom in a way that feels a bit crowded.
Verdict: While FLUX.2 [dev] Turbo creates a more visually striking graphic design with vibrant accents, it fails significantly on content, using gibberish text and placing pizzas under the appetiser section. GPT Image 1.5 provides a fully functional, highly professional menu with accurate English text, realistic pricing, and a logical layout that matches the prompt's structural requirements better.
Adorable Baby Animals in Sunny Meadow
Text-to-Image“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Perfect adherence to all four requested animal species.
- + Excellent dynamic composition showing interaction and movement.
- + Clearer rendering of individual fur strands and environmental dew sparkles.
- − The bunny is slightly static compared to the other playful animals.
GPT Image 1.5
- + Beautiful warm lighting with strong rays and cohesive color palette.
- + Emphasizes the 'expressive eyes' part of the prompt well.
- + High level of cute, wholesome appeal.
- − The fox kit's anatomy is slightly distorted with its paw placement.
- − Less defined action/play compared to the first image.
Verdict: Both models captured the requested atmosphere and lighting beautifully. FLUX.2 [dev] Turbo followed the prompt more precisely by clearly rendering all four distinct animal species with excellent fur detail and a sense of 'tumbling' motion, whereas GPT Image 1.5 felt slightly more cluttered in its arrangement.
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
FLUX.2 [dev] Turbo
- + Captures a very wide variety of butterflies and orchid colors.
- + The iron framework is symmetrical and meticulously detailed.
- + Effective use of dew drops on the foreground leaves.
- − Butterflies look somewhat 'pasted on' with flat lighting compared to the environment.
- − The mist feels a bit more like a flat haze than a volumetric atmosphere.
GPT Image 1.5
- + Superior lighting with beautiful volumetric sunbeams (crepuscular rays) and caustic-like highlights.
- + The architecture shift to a circular dome creates a more immersive and grand composition.
- + Better integration of the butterflies into the lighting and depth of the scene.
- − Fewer butterflies than requested, though they are higher quality.
- − The water/mist effect on the lens slightly obscures fine details in the background.
Verdict: While FLUX.2 [dev] Turbo provides a very clean and symmetrical layout with a high quantity of butterflies, GPT Image 1.5 wins on sheer visual atmosphere. GPT Image 1.5 correctly interprets the 'caustics' and 'misty atmosphere' prompts to create a stunning, god-ray filled interior that feels more like a 8K masterpiece than the flatter lighting of FLUX.2.
Heroic Super Hero Portrait
Text-to-Image“Hyper-photorealistic full-body portrait of a female superhero standing triumphantly on a New York skyscraper rooftop at golden sunset, wearing a classic modest superhero costume with flowing cape, chest emblem, gloves, and boots in red and blue colors, practical design, short hair, strong determined heroic expression looking into the distance, powerful confident stance with hands on hips and cape billowing dramatically in the wind, detailed urban cityscape background, warm natural sunlight with sharp shadows and fabric highlights, ultra-sharp textures on suit, hair, and concrete, 8K masterpiece, empowering family-friendly style.”
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent adherence to the 'modest' attire request with full leggings.
- + Very realistic lighting integration, with the sun correctly positioned as a back-light source.
- + Superior skin texture and hair detail that feels photorealistic.
- − The placement of the hands on the hips looks slightly anatomically stiff.
- − The New York City layout is a bit generic compared to the recognizable landmarks in the other image.
GPT Image 1.5
- + Features iconic NYC landmarks like the Empire State Building and Chrysler Building.
- + Strong, dynamic cape billowing effect.
- + Accurately captures the 'short hair' requirement with a pixie cut.
- − Failed the 'modest' requirement by generating a short skirt/tunic instead of full-body coverage.
- − The lighting on the character's face doesn't quite match the intensity of the sunset behind her.
- − The proportions of the legs seem slightly exaggerated compared to the torso.
Verdict: FLUX.2 [dev] Turbo followed the prompt's instruction for a 'modest' costume much better than GPT Image 1.5, which opted for a more traditional short-skirted Supergirl look. FLUX.2 also achieved a more convincing photorealistic finish with superior lighting and texture, although GPT Image 1.5 captured a more iconic New York skyline.
Intricate Floral Mandala
Text-to-Image“Perfectly symmetrical mandala made entirely of real flowers, petals, leaves, fruits, and seeds in vibrant natural colors, intricate layered patterns with radial symmetry, top-down view on a soft neutral background, hyper-detailed organic textures and subtle shadows, photorealistic, 8K masterpiece.”
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Sophisticated photorealistic textures with natural depth and soft shadows.
- + Excellent variety of organic materials including small fruits and visible seeds.
- + Very high level of detail in the petal veins and flower centers.
- − Slightly less geometric precision compared to Model B, with some organic irregularities.
- − Center focal point is a bit less defined than the outer layers.
GPT Image 1.5
- + Exceptional geometric symmetry and precise alignment of elements.
- + Vibrant, high-contrast color palette that makes the pattern pop.
- + Clear inclusion of acorns and berries providing distinct textural variety.
- − Lighting is very flat and uniform, losing the depth found in a real top-down photograph.
- − Some elements appear slightly repetitive or digitally perfect rather than organic.
Verdict: FLUX.2 [dev] Turbo produces a more photorealistic result with beautiful soft lighting and natural textures that feel like a physical installation. GPT Image 1.5 offers superior mathematical symmetry and a more colorful layout, but feels flatter and more like a digital illustration than a real-world photograph. FLUX.2 [dev] Turbo is preferred for its convincing 8K organic textures and depth.
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 [dev] Turbo
- + Perfect text rendering including the grave accent in 'Caffè'.
- + Sophisticated vector emblem style with a beautiful aged paper background texture.
- + Excellent composition and balance between the cloche, text, and banner.
- − The steam lines are a bit thick compared to the overall minimalist aesthetic.
GPT Image 1.5
- + Dynamic and creative typography for the main brand name.
- + Good use of shading and stippling on the cloche dome to create a retro feel.
- + Follows all prompt elements including the 'Est. 1720' banner.
- − The background is plain white, ignoring the 'light background with subtle texture' request.
- − The banner illustration is a bit clunky with inconsistent line weights.
- − Missed the accent mark on 'Caffè'.
Verdict: FLUX.2 [dev] Turbo produced a superior logo that perfectly captured the 'vintage minimalist' aesthetic with high-quality vector-style execution and a well-textured background. GPT Image 1.5 had interesting typography but failed to include the texture requested in the prompt and lacked the refined finish of the first model.
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 [dev] Turbo
- + Excellent text rendering with accurate names and secondary details.
- + Complex layout that includes extra information like astronaut silhouettes.
- + Adheres well to the navy and white palette requested.
- − The layout is cluttered and confusing, making the sequence of steps hard to follow.
- − The icons are more illustrative and less 'flat-vector' than requested.
- − Includes text errors like 'Saturn licon' and duplicate labels.
GPT Image 1.5
- + Perfect adherence to the flat-vector style with crisp lines and a structured layout.
- + Very clear, logical progression through the mission steps.
- + Strictly follows the NASA-inspired color palette for a professional look.
- − The top header text is cut off partially.
- − Slightly less detailed rocket and moon icons compared to the other model.
Verdict: GPT Image 1.5 is the clear winner as it perfectly captures the 'clean, modern vector infographic' aesthetic with a logical, easy-to-read vertical layout. While FLUX.2 [dev] Turbo manages impressive text and detail, its composition is messy and lacks the professional infographic feel requested in the prompt.
FLUX.2 [dev] Turbo
Distilled version of Black Forest Labs' FLUX.2 [dev] outperforming it at a cheaper price. Developed by fal.ai.
GPT Image 1.5
OpenAI's state-of-the-art image generation model with better instruction following and adherence to prompts