Black Forest Labs' aesthetically-tuned 12-billion parameter flow transformer optimized for high-quality images with incredible aesthetics, suitable for personal and commercial use
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.1 Krea [dev]
#47 of 62 in Text-to-Image
FLUX.2 [klein] 4B
#32 of 62 in Text-to-Image
Where the votes landed
FLUX.1 Krea [dev]
0%
win rate
Ties
0%
FLUX.2 [klein] 4B
0%
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
FLUX.1 Krea [dev]
- + Excellent rendering of glass reflections and refraction
- + Highly realistic lighting and depth of field
- + Accurate depiction of a glass cube with a thick base
- − The sphere appears slightly dark and opaque rather than a clear blue
- − The background is very dark, making the plant harder to distinguish
FLUX.2 [klein] 4B
- + Perfectly matches the blue sphere description with a saturated tone
- + Clear and bright implementation of soft window light from the left
- + Composition feels balanced and clean
- − The red book is unnaturally thin for its dimensions
- − The glass cube has less realistic refractive properties compared to Model A
Verdict: Both models followed the prompt perfectly, placing all elements in the correct relational positions. FLUX.1 Krea (dev) produced a more photographically realistic image with sophisticated light and reflections, whereas FLUX.2 [klein] 4B followed the color descriptions more vividly but struggled slightly with the proportions of the book.
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.1 Krea [dev]
- + Excellent handling of motion blur on background vehicles
- + Very convincing wet pavement reflections
- + Follows the 'imperfect framing' prompt well with a lower camera angle
- − Anatomical issues with hands and how they grip the bike
- − The bike's frame geometry is slightly warped near the bottom bracket
FLUX.2 [klein] 4B
- + Stronger emotional narrative and focus on the subject
- + Higher level of fine detail in the man's clothing and the bicycle's mechanical parts
- + Beautiful bokeh and lighting from the car headlights
- − Cars in the background are static rather than having the requested motion blur
- − The bicycle has structural inconsistencies, such as the seat post merging into the frame
Verdict: FLUX.1 Krea [dev] does a better job of capturing the specific cinematic technical requirements like motion blur and imperfect street photography framing. However, FLUX.2 [klein] 4B produces a much more detailed subject with better skin textures and lighting, even if it ignores the motion blur request. FLUX.1 Krea [dev] is the winner for adhering more strictly to the complex atmospheric cues of the prompt.
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.1 Krea [dev]
- + Excellent high-contrast cinematic lighting with realistic reflections on the metal armor
- + Intricate, fine engravings on the plate armor that look authentic
- + Better adherence to the 'battle-worn' descriptor with visible grime and dried blood
- − The beads in the hair are very small and blend into the highlights
- − Shadows are quite deep, obscuring some of the leather strap detail
FLUX.2 [klein] 4B
- + Clearer representation of the braided hair with colorful small beads
- + The torchlight sources are physically visible in the background which adds context
- + Sharp textures on the cloth underlayer and leather straps
- − The lighting feels flatter and less cinematic compared to Image A
- − The armor engravings look slightly more stylized and less realistic
- − The character looks less 'battle-worn' and more 'cleanly scarred'
Verdict: FLUX.1 Krea [dev] produces a much more atmospheric and cinematic result that captures the 'battle-worn' essence through moody lighting and gritty skin textures. While FLUX.2 [klein] 4B followed the 'beads' and 'leather straps' prompts more literally with higher visibility, it lacks the photographic realism and intricate metalwork detail found in the first image.
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.1 Krea [dev]
- + Excellent structure that mimics a real physical menu flyer.
- + Strong adherence to the requested headers like 'Appetizers' and 'Maines'.
- + Clean use of white space and professional bold sans-serif fonts.
- − Repeats the 'Appetizers' header twice instead of using 'Pizza'.
- − Text contains minor gibberish/artifacts in the smaller body copy.
FLUX.2 [klein] 4B
- + Highly creative grid-based layout that feels very modern.
- + Excellent food photography quality with vibrant colors.
- + Good use of bold logo-style branding and accents.
- − The text is largely unreadable with strange character artifacts (e.g., 'MDANS', 'FRAZZA').
- − The layout is a bit scattered and less functional for a real menu.
Verdict: FLUX.1 Krea [dev] is the superior choice because it produces a functional, professional menu layout that feels realistic and follows the hierarchy of information expected in a restaurant setting. While FLUX.2 [klein] 4B has higher visual appeal in its food photos, its text rendering is significantly worse and the layout is less practical for a template.
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.”
AI Judge Analysis
FLUX.1 Krea [dev]
- + Excellent text rendering with no spelling errors
- + Follows the 'exploded' instruction by showing visible space between all components
- + Great integration of the 'fiery background' with actual coals and flames
- − The price is not inside a starburst as requested
- − The layout is a bit cluttered with small gibberish text at the bottom
FLUX.2 [klein] 4B
- + Good use of the starburst effect for the price
- + Vibrant colors and a strong sense of motion through embers
- + High visual quality on the burger textures
- − Major text failures on the primary title ('MAAC AGID BUIRGER')
- − Fails to create an 'exploded' view; the burger is mostly assembled
- − Ignores the request for a fiery/glowing effect on the main text
Verdict: FLUX.1 Krea [dev] is the clear winner because it successfully rendered all requested text accurately and adhered to the 'exploded burger' composition. FLUX.2 [klein] 4B failed significantly on the text rendering (garbled characters) and provided a standard burger rather than an exploded one.
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.”
AI Judge Analysis
FLUX.1 Krea [dev]
- + Excellent chalk-like aesthetic with realistic dust and smudge textures.
- + Elegant cursive handwriting that matches the requested style profile.
- + Clean composition and professional framing of the chalkboard.
- − Significant spelling errors and hallucinations in the text, such as 'Gruffle' and 'Clvrucuations'.
- − The text transitions into a very clean, almost digital-looking font at the bottom.
- − Failed to accurately complete the specific menu item list requested in the prompt.
FLUX.2 [klein] 4B
- + Followed the specific menu item text much more closely than the competitor.
- + Captured a more natural, irregular chalk texture consistent across the entire board.
- + Successfully rendered complex ingredients like 'Lemon & Herbs' and the 'Brown Butter' item.
- − Several minor spelling typos such as 'Truffel', 'Musheram', and 'Ootrpous'.
- − The spacing in the title is awkward with a large gap in 'S PECIALS'.
- − Overall lower resolution and slightly less refined handwriting compared to the other model.
Verdict: FLUX.1 Krea [dev] produces a much more visually polished and aesthetically pleasing cafe scene, but it fails significantly on text accuracy and prompt adherence regarding the specific menu items. FLUX.2 [klein] 4B follows the prompt's instructions for the menu list much better and maintains a consistent chalk texture throughout, even though it suffers from typical AI spelling phoneticisms. FLUX.2 [klein] 4B is the better choice for this specific text-heavy task despite the lower image resolution.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
AI Judge Analysis
FLUX.1 Krea [dev]
- + Successfully followed the difficult logic of having the horse 'on top' by merging it into the suit/craft.
- + High cinematic quality with dramatic lighting and a clear view of Earth.
- + Realistic textures on the spacesuit and horse hair.
- − The composition is physically confusing, creating a hybrid creature rather than a separate horse and astronaut.
- − The horse lacks ears, affecting the realism of the animal model.
FLUX.2 [klein] 4B
- + Very high detail in the star fields and the rendering of the horse's coat.
- + Clear, high-resolution textures on the astronaut's suit and visor.
- + Good handling of the 'walking' motion over the planet surface.
- − Failed the primary prompt constraint of horse on top; instead, the astronaut is riding the horse.
- − The shadow on the clouds below does not match the horse's light source naturally.
Verdict: This challenge featured a specific logical constraint: 'horse on top, not vice versa.' FLUX.1 Krea (dev) attempted this by creating a surreal hybrid where the horse occupies the upper body and head area of a robotic suit, whereas FLUX.2 [klein] 4B completely ignored the instruction and produced a standard astronaut-riding-horse image. Despite the anatomical oddities in Image A, it is the clear winner for attempting the complex prompt logic.
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.”
AI Judge Analysis
FLUX.1 Krea [dev]
- + Excellent texture on the capybara's fur
- + Good use of cinematic lighting and bokeh
- + Includes the specific 'TAXI' text on the cap's front
- − Includes an unnecessary second human passenger in the front seat
- − Capybara only has one paw clearly on the steering wheel instead of both
- − Composition is a bit far away, hiding some interior detail
FLUX.2 [klein] 4B
- + Perfect adherence to the 'both paws on steering wheel' instruction
- + Composition creates a more intimate and believable taxi interior feel
- + The businesswoman's expression and coat perfectly match the prompt requirements
- − The text on the hat is illegible/gibberish
- − The capybara's right eye/side of face looks slightly flat due to the lighting
- − Slightly less 'photorealistic' fur detail compared to the other model
Verdict: FLUX.2 [klein] 4B is the winner because it followed all specific prompt instructions, including the number of passengers and the specific placement of the capybara's paws. While FLUX.1 Krea [dev] has slightly more consistent fur rendering, it failed the core instruction of having a single female passenger in the back seat by adding a second human in the front.
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.”
AI Judge Analysis
FLUX.1 Krea [dev]
- + Excellent decorative border incorporating webs and thorns as requested.
- + Strong graphic design layout with good contrast.
- + Incorporates all requested text elements, including the specific date and address.
- − Noticeable spelling errors in the title ('Pasty Halloween') and bottom banner ('You are are', 'Tright').
- − The date is oddly split and partially obscured in the center.
- − The composition feels slightly cluttered around the bottom scroll.
FLUX.2 [klein] 4B
- + Highly atmospheric moody night sky and cinematic lighting.
- + Clean and legible text at the bottom for the date and location.
- + Consistent gothic art style across all elements including the background silhouettes.
- − Major spelling errors in the main title text ('Hallbwom Party niisation').
- − The border is slightly less detailed than the competitor's version.
- − Missed the specific 'Time' detail in the bottom text.
Verdict: FLUX.1 Krea (dev) followed the complex layout instructions better, including the thorn/web border and all specific event details, although it suffered from several significant typos. FLUX.2 [klein] 4B produced a much more visually cohesive and cinematic image with a superior background, but the severe misspelling of 'Halloween' makes it less functional as an invitation. FLUX.1 Krea (dev) is the winner for better prompt adherence regarding the specific text fields and border details.
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.1 Krea [dev]
- + Perfectly executed the 'small raised diorama base' request.
- + Correctly rendered all text and included the accurate Japanese flag icon.
- + Achieved a cleaner, more professional 'isometric miniature' aesthetic with distinct lighting.
- − The salmon texture is slightly more plasticky compared to Model B's rendering.
FLUX.2 [klein] 4B
- + Excellent PBR materials on the salmon and ceramic plate showing realistic gloss and translucency.
- + Very clean, centered composition.
- − Failed to include the 'small raised diorama base' as part of the scene architecture.
- − Spelling error in the text ('SUSH' instead of 'SUSHI') and an incorrect flag representation.
- − Lack of the 45-degree isometric cube پایه specified in the prompt.
Verdict: FLUX.1 Krea [dev] followed the prompt instructions much more accurately, successfully incorporating the diorama base, correct text, and the Japanese flag. While FLUX.2 [klein] 4B had slightly superior material textures for the sushi itself, it failed significantly on technical requirements like spelling and specific layout elements.
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.1 Krea [dev]
- + Excellent fur texture with high fine-detail rendering
- + Dynamic variety in poses including a jumping motion
- + Very strong interpretation of the 'god rays' lighting prompt
- − Failed to include a baby bunny, substituting it with a second kitten
- − The fox kit has slightly unusual dark paws that look more like a cat's
FLUX.2 [klein] 4B
- + Beautifully rendered background with lush, colorful wildflowers
- + Clearer eyes and expressive facial features on the animals
- + The lighting feels very warm and natural, hitting the fur edges nicely
- − Also failed to include the requested baby bunny, providing two kittens instead
- − The butterfly on the right has slightly messy wing anatomy
Verdict: Both models failed to include all four requested animals, with both FLUX.1 Krea [dev] and FLUX.2 [klein] 4B omitting the baby bunny in favor of a second kitten. FLUX.1 Krea [dev] is the winner because its rendering of fur and the 'god rays' is more technically impressive and fits the '8K masterpiece' requirement better, despite the slightly less colorful flower field.
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.1 Krea [dev]
- + Ornate and authentic vintage woodcut illustration style.
- + Includes all elements like the steam, cloche, and banner correctly.
- + Exceptional texture and balance for a retro emblem.
- − Misspelled 'Florian' as 'FLANDR'IN'.
- − Text layout is a bit cluttered around the banner.
FLUX.2 [klein] 4B
- + Cleaner, more minimalist vector style.
- + Correct 'Caffè' spelling with the accent mark.
- + Very clean background texture.
- − Misspelled 'Florian' as 'FLAXTION'.
- − Redundant repetition of 'Est. 1720' both on and below the banner.
- − The cloche illustration is very basic compared to the requested vintage style.
Verdict: FLUX.1 Krea [dev] captures the 'vintage' and 'cloche dome' aesthetic much more effectively with high-quality line work, though it struggles with the text spelling. FLUX.2 [klein] 4B is cleaner but fails on the prompt's specific naming and includes redundant text, making it look less like a professional logo.
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.1 Krea [dev]
- + The header text 'Apollo 11' is perfectly rendered and legible.
- + Successfully adhered to the NASA-inspired dark navy and white palette.
- + The flat vector style is consistent across the icons.
- − Failed to provide the specific icons requested, showing astronauts instead of a Saturn V or descent stages.
- − The logical flow of the infographic steps is non-sensical and disorganized.
- − Includes random planets with rings instead of focusing on Earth and the Moon.
FLUX.2 [klein] 4B
- + Successfully followed the specific sequence of icons requested including the Saturn V and Lunar Module.
- + Better infographic layout with labeled steps corresponding to the mission phases.
- + Maintains a clean, professional vector aesthetic that matches the 'modern infographic' prompt.
- − The text rendering is very poor with numerous spelling errors in both the title and labels.
- − The color palette leans a bit too much into tan/brown rather than the requested navy/light gray.
Verdict: While FLUX.1 Krea (dev) has superior text rendering for the main title, it completely fails to follow the content requirements of the prompt, producing generic space icons that don't represent the Apollo mission steps. FLUX.2 [klein] 4B followed the complex instructional steps much better, accurately representing the Saturn V, Earth, trajectory, and Lunar Module in a logical sequence, making it the better infographic despite the garbled text.
Explore each model
Black Forest Labs' compact, open-source image generation model with sub-second inference, optimized for production and near real-time applications with multi-reference support