Black Forest Labs' 12-billion parameter flow transformer for high-quality text-to-image generation, suitable for personal and commercial use with streaming support
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.1 [dev]
24.6
arena score
#16 of 62 in Text-to-Image
Skill signature
· Text-to-Image
FLUX.1 [schnell] FP8
19.0
arena score
#46 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.1 [dev]
100.0%
win rate
Ties
0.0%
FLUX.1 [schnell] FP8
0.0%
win rate
100.0%
0.0% ties
0.0%
Shared challenges
13
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent depiction of physics with the blue sphere appearing to float or be suspended within the glass mass.
- + High-quality textures on the red leather book.
- + Sophisticated lighting and caustic effects on the wooden table.
- − The 'cube' is more of a rectangular prism.
- − The plant is almost entirely behind the cube rather than being clearly visible through it.
FLUX.1 [schnell] FP8
- + Successfully shows the plant through the glass as requested.
- + The sphere is well-detailed and behaves realistically with the internal shelf.
- + Good adherence to the window lighting requirement.
- − The glass structure is a tall rectangular prism rather than a cube.
- − The internal shelf was not requested and complicates the simplicity of the prompt.
- − The reflection of the book on the internal shelf looks slightly metallic/opaque.
Verdict: Both models struggled to generate a literal cube, opting for rectangular prisms instead. FLUX.1 [dev] produced a more artistic and physically interesting image with a floating sphere effect, while FLUX.1 [schnell] FP8 better followed the specific instruction to show the plant through the glass. FLUX.1 [dev] is the winner due to superior texture rendering and a more aesthetically pleasing composition.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent depiction of rain with visible droplets and wet textures on the jacket.
- + High visual realism in skin texture and clothing materials.
- + Effective use of shallow depth of field and motion blur on the passing car.
- − The framing is a bit centered for a 'candid' street photo.
- − The bicycle proportions look slightly small relative to the man.
FLUX.1 [schnell] FP8
- + Dynamic composition with a strong 'street photo' feel due to the angle and background elements.
- + Beautiful reflections on the wet pavement.
- + Accurate red bicycle color as requested.
- − The man is wearing a short-sleeved shirt, which feels inconsistent with the rainy weather.
- − Lack of visible rain droplets compared to Model A.
- − Minor anatomical wonkiness with the hands on the handlebars.
Verdict: FLUX.1 [dev] captures the atmosphere of a rainy day much more convincingly, showing moisture on the man's jacket and clear rain streaks. While FLUX.1 [schnell] FP8 offers a more interesting street composition, it fails to depict the man dressed appropriately for the weather and lacks the fine skin details present in the other model.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent depiction of the requested beads in the braids
- + High-quality skin texture and lifelike eye details
- + Better representation of 'ornate engraved plate armor'
- − Character looks more youthful and clean rather than 'battle-worn'
- − Scars and dirt are very faint, almost resembling freckles
FLUX.1 [schnell] FP8
- + Stronger 'battle-worn' aesthetic with intense expression and rugged texture
- + Excellent capture of warm torchlight reflecting off skin and hair
- + Includes detailed leather straps and cloth underlayers as requested
- − The beads in the hair are less prominent and look like simple metallic rings
- − Armor detail is slightly more abstract and less 'engraved plate' than Image A
Verdict: Both models followed the prompt well, but FLUX.1 [schnell] FP8 (Image B) captured the 'battle-worn' mood and lighting much more effectively through the character's expression and skin texture. FLUX.1 [dev] produced a cleaner, more aesthetically pleasing portrait with better armor engraving, but it missed the grit of the prompt.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Features a very clean, professional minimalist layout typical of modern graphic design.
- + High-quality photographic elements with realistic lighting and shadows.
- + Effective use of negative space and clear hierarchy between headings and body text.
- − Failed to include a 'pizza' specific section and the overall grid for photos.
- − The main body text is mostly gibberish despite the header being clear.
FLUX.1 [schnell] FP8
- + Adhered better to the 'grid' request by organizing several food photos in a structured layout.
- + Included the requested sections for Appetizers (Apptizers), Pizza (Pizzal), and Mains.
- + Text density and placement more closely resemble a functional multi-page or bifold menu.
- − The visual quality of the food items is slightly lower/muddier compared to Image A.
- − Several spelling errors in headings such as 'ORCETERS', 'APPTIZERS', and 'PIZZAL'.
- − The 'MENU' title at the top is poorly integrated and uses a serif font, contrary to the 'sans-serif' request.
Verdict: FLUX.1 [dev] produced a much more professional and aesthetically pleasing design that looks like high-end graphic work, though it missed some specific layout instructions. FLUX.1 [schnell] FP8 followed the structural prompt requirements more closely (sections and grid), but suffered from numerous typos and a slightly less refined visual style. FLUX.1 [dev] is the preferred choice for its superior clarity and realistic professional execution.
Magic Burger Explosion: Fiery Photorealism Challenge
Text-to-Image“Ad for 'Magic Burger'. Dynamic, exploded burger with all components (bun, patty, cheese, lettuce, tomato, sauce) suspended in mid-air. Emphasize photorealistic detail and a sense of motion. Dark, fiery background with glowing embers. Integrate text: 'MAGIC BURGER' as a prominent title, 'LIMITED TIME ONLY' as a secondary message, and '€6.99' in a starburst, all rendered with a fiery, glowing effect.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Successfully rendered clean, legible text for the secondary message and price.
- + Captures a very professional, high-end food photography aesthetic.
- + The lighting and fiery base create a cohesive and atmospheric dark background.
- − Completely missed the 'MAGIC BURGER' title text.
- − The burger stack is very tall and repetitive, looking less like a single burger exploded and more like multiple burgers stacked.
- − Failed to include the starburst element for the price.
FLUX.1 [schnell] FP8
- + Included the 'MAGIC BURGER' title text as requested.
- + The 'exploded' effect is much more dynamic and energetic with ingredients flying outwards.
- + Included the requested starburst element for the price.
- − Significant text rendering errors including 'LIIMITED' and 'NEEY'.
- − Incorrect price rendering, showing '€69' and '€.99' separately instead of '€6.99'.
- − The text lacks the requested 'fiery, glowing effect' and looks like a flat graphic overlay.
Verdict: FLUX.1 [dev] produced a more polished and photorealistic image with accurate spelling, but it failed to include the primary brand name. FLUX.1 [schnell] FP8 captured the specific layout requests much better, including the title and starburst, but suffered from severe typos and poor text integration. Overall, FLUX.1 [dev] is preferred for its visual quality, though it requires further prompting to include the missing title.
Chalkboard Menu
Text-to-Image“Handwritten-style chalkboard menu in a cozy café, all text rendered in the exact same realistic chalk handwriting style with natural variations in letter size, slight slant, and chalk texture — no printed or digital fonts anywhere on the board. Title at the top in elegant cursive chalk handwriting: ‘TODAY’S SPECIALS – APRIL 30, 2026’. Below it, three menu items also in the same handwritten chalk style: ‘Truffle Mushroom Risotto – $24’, ‘Grilled Octopus with Lemon & Herbs – $28’, ‘Brown Butter Chocolate Chip Cookies – $9’. At the very bottom, smaller text in the identical handwritten chalk style (slightly smaller but still clearly legible with the same handwriting characteristics): ‘All items made fresh daily • Ask about our gluten-free options’. Warm ambient café lighting, visible chalk dust and smudges, realistic handwriting imperfections, no clean printed text anywhere.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent text accuracy for all requested menu items and numbers.
- + Very realistic chalk texture with dusty edges and pressure variations.
- + Well-composed layout that looks like a professional cafe chalk board.
- − Slight spelling errors in the bottom disclaimer text (e.g., 'drish' instead of 'fresh').
- − Overall handwriting is slightly too uniform, bordering on font-like in certain sections.
FLUX.1 [schnell] FP8
- + Good natural variation in handwriting style and slant.
- + Correct title text and date rendering.
- − Significant spelling and coherence issues in the menu items (e.g., 'Octemon', 'Buthowith').
- − Repeated text lines and messy layout compared to the prompt requirements.
- − Incorrect pricing and disorganized placement of dollar amounts.
Verdict: FLUX.1 [dev] is the clear winner as it successfully rendered almost all the requested text with high legibility and a convincing chalk texture. FLUX.1 [schnell] FP8 struggled with the complexity of the menu, resulting in multiple spelling errors, redundant lines of text, and a messy overall composition.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent anatomical rendering of both the horse and the astronaut.
- + Clean, cinematic lighting and high-quality textures.
- + Strong composition with a clear focal point against the Earth's horizon.
- − Failed to follow the negative prompt instruction 'horse on top, not vice versa'.
- − Standard interpretation lacking the requested surrealism.
FLUX.1 [schnell] FP8
- + Successfully attempted the surrealist logic of 'horse on top' of the equipment.
- + Strong cinematic composition with a large planetary backdrop.
- − Coherence issues where the horse/astronaut hybrid becomes a second horse head.
- − Anatomical errors in the horse's legs and the floating mechanical structure.
- − Loss of the 'astronaut' character in favor of a robotic equipment block.
Verdict: While FLUX.1 [schnell] FP8 attempted the specific 'horse on top' logic requested in the prompt, it resulted in a confused composition with significant anatomical glitches and the disappearance of the astronaut character. FLUX.1 [dev] produced a much higher quality, aesthetically pleasing image, even though it defaulted to the traditional 'astronaut on horse' arrangement.
The Capybara Taxi Driver
Text-to-Image“Photorealistic scene inside a yellow New York taxi at night. A capybara is driving, wearing a yellow taxi driver cap and a dark jacket. It has a calm, professional expression and both front paws on the steering wheel. In the back seat sits a human businesswoman in a coat, looking at her phone with a completely normal, bored expression (as if this is just another normal ride). Through the windows you can see the streets of Manhattan at night with blurred lights. Realistic taxi interior, photorealistic, detailed fur and fabric, 35mm lens, night lighting with reflections, shallow depth of field.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent photorealistic texture on the capybara's fur
- + Accurate execution of the passenger sitting in the back seat
- + Strong cinematic lighting and Depth of Field
- − The passenger appears to be in the front passenger seat rather than the back seat
- − The steering wheel placement is slightly obscured by the foreground
FLUX.1 [schnell] FP8
- + Successfully places the passenger clearly in the back seat
- + Accurate taxi hat branding on the capybara
- + Good background city details
- − The passenger has a distorted third hand holding a second phone
- − The capybara's paws are not both on the steering wheel as requested
- − Lower overall photorealism compared to Model A
Verdict: FLUX.1 [dev] produced a much more realistic and high-quality image with better lighting and fur textures, although it placed the passenger in the front seat. FLUX.1 [schnell] FP8 correctly placed the passenger in the back, but suffered from significant anatomical errors (three hands) and failed the specific instruction to have both paws on the steering wheel.
The Halloween Invitation
Text-to-Image“Vintage gothic Halloween party invitation. Dark parchment poster, spooky border with webs and thorns, central glowing jack-o-lantern, bats, twisted trees, moody night sky. Add elegant gothic title text saying "Halloween Party Invitation", a small scroll banner saying "You are invited to a night of frights", and event details at the bottom: Date: 30.10.2026 Time: 7pm Location: The Arches, NYC Spooky but polished, cinematic lighting, square format.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
100% wins
0% ties
0% wins
AI Judge Analysis
FLUX.1 [dev]
- + Strong composition with a clean, thorn-based border.
- + Excellent lighting on the central jack-o-lantern.
- + The lower body of the text is highly legible and follows the prompt's details closely.
- − The main title text in the banner is misspelled and distorted.
- − Repetition of information in the 'Time' field.
FLUX.1 [schnell] FP8
- + Captures a better 'vintage parchment' texture and atmosphere.
- + Includes many bats and twisted trees that blend well with the background.
- + Main title 'HALLOWEEN INVITATION' is very clear and stylistically appropriate.
- − Text rendering is very poor in the body including numerous typos like 'Tine' and 'Theaches'.
- − The central pumpkin feels less integrated into the scene's lighting compared to Model A.
- − Grammar and spelling on the scroll banners are incorrect.
Verdict: FLUX.1 [dev] produces a much cleaner and more professional-looking layout with superior lighting, although it struggles with the primary banner text on the scroll. FLUX.1 [schnell] FP8 handles the background atmosphere and large title text better, but the excessive character errors in the event details and banners make it unusable for an actual invitation. FLUX.1 [dev] is the preferred choice for its overall polish and better execution of the prompt's complex text requirements.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent PBR textures with a soft, clay-like miniature feel
- + Accurate 45-degree isometric perspective
- + Includes a clean flag icon and bold primary text
- − Text contains typos like 'SUSH CATON' instead of 'SUSHI'
- − Includes extra floral garnish not explicitly requested
FLUX.1 [schnell] FP8
- + Features a variety of sushi types (nigiri and maki)
- + Ultra-clean visual aesthetic with high clarity
- + Strong adherence to the 'cartoon scene' and 'no garnish' request
- − Failed to include the word 'SUSHI' in text
- − Text is repetitive 'JAPAN JAPAN' with a fused flag icon
- − Lighting is a bit flat compared to Model A
Verdict: FLUX.1 [dev] produced a more sophisticated miniature look with superior material rendering (PBR) and lighting, despite some minor typos in the small text. FLUX.1 [schnell] FP8 captured the cartoon style and variety of sushi well but failed to render the text 'SUSHI' and had a less convincing diorama feel. FLUX.1 [dev] is the winner for its professional composition and texture quality.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent character separation and clarity.
- + Whimsical, storybook-like aesthetic that fits the 'joyful' vibe.
- + High-quality rendering of fur and lighting effects.
- − Failed to include a tabby kitten, providing two dogs instead.
- − The style leans towards 3D animation/CGI rather than 'hyper-photorealistic'.
- − Anatomically awkward limbs on the fox/bunny-like characters.
FLUX.1 [schnell] FP8
- + Successfully included a golden retriever, kittens, and a fox kit.
- + Texture is very detailed with convincing 'ultra-detailed soft fur'.
- + Captures the 'tumbling together' aspect of the prompt more effectively.
- − Missed the baby bunny entirely.
- − Composition is a bit cluttered with overlapping characters.
- − Visible AI artifacts around the small kitten's paws and the fox's face.
Verdict: Both models struggled to include all four specific animals, but FLUX.1 [schnell] FP8 came closer to the prompt by including the kitten and fox kit. While FLUX.1 [dev] produced a cleaner, more balanced image, its failure to generate the requested animals and its move toward a cartoonish style makes FLUX.1 [schnell] FP8 the preferred choice for prompt adherence and texture realism.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Features a clean, vector-emblem style layout.
- + Correctly implements the requested 'Est. 1720' text.
- + Captures the classic vintage typography aesthetic well.
- − Misspells the main name as 'Flarilaan' and 'Reseaurant'.
- − Includes extraneous, nonsensical numbers like 11011 and 1941.
FLUX.1 [schnell] FP8
- + Creative integration of the cloche dome as a central architectural element.
- + Good use of the requested banner and warm color palette.
- − Significant spelling errors including 'Afe Flamilan'.
- − The typography on the banner is poorly aligned and cluttered.
- − The steam effects look like disconnected swirls rather than smoke.
Verdict: Both models struggled with the specific spelling of 'Caffè Florian', but FLUX.1 [dev] produced a much more professional and balanced vector logo that fits the 'minimalist' request. FLUX.1 [schnell] FP8's design is messy, with poorly integrated text and a cluttered composition that fails the minimalist requirement.
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.”
FLUX.1 [dev]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [dev]
- + Excellent vector aesthetic with a clean, centered composition.
- + Superior adherence to the specific 'navy, white, muted red' color palette.
- + Includes a bottom legend that attempts to categorize the requested mission steps.
- − Text rendering is mostly gibberish compared to the other model.
- − The icons for planets include rings (like Saturn) which are not relevant to the Apollo 11 mission.
FLUX.1 [schnell] FP8
- + Stronger text rendering with legible headers for mission steps.
- + Cleaner iconography that feels more informative and structured.
- + Includes a clear lunar module illustration that fits the theme well.
- − Composition is slightly disjointed with icons split into two different sections.
- − Color palette is more limited than Model A, lacking the 'muted red' accents in the main icons.
Verdict: FLUX.1 [dev] produced a more visually cohesive poster with a higher-quality 'vector' feel, though its text is unreadable. FLUX.1 [schnell] FP8 was more successful at following the specific steps requested in the prompt with legible labels, even though the layout was less elegant overall. FLUX.1 [schnell] FP8 is preferred for its informative clarity, while FLUX.1 [dev] is superior for purely aesthetic graphic design.
Next steps
Explore each model
FP8 quantized variant of Black Forest Labs' FLUX.1 [schnell] model, offering ~2x faster inference with reduced precision while maintaining high-quality image generation in 4 steps