Black Forest Labs' premium multimodal flow transformer with greatly improved prompt adherence and typography generation for in-context image generation and editing without compromise on speed
Settled by community votes across 17 shared challenges, with an AI judge weighing in on each.
FLUX.1 Kontext [max]
#21 of 62 in Text-to-Image
FLUX.2 [klein] 4B
#32 of 62 in Text-to-Image
Where the votes landed
FLUX.1 Kontext [max]
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 Kontext [max]
- + Excellent handling of caustics and light refraction through the glass cube
- + Rich textures on the wooden table and red book cover
- + Strong adherence to the lighting and placement requirements
- − The blue sphere has an unusual glittery/sandy texture that wasn't requested
FLUX.2 [klein] 4B
- + Smooth and clean rendering of the blue sphere
- + Good composition and natural-looking plant in the background
- − The glass cube lacks realistic reflections and refraction
- − The book lacks the detail and depth seen in the competing model
Verdict: FLUX.1 Kontext [max] produced a superior image with highly realistic lighting, including intricate caustics on the table and sophisticated refractions of the plant through the glass. While FLUX.2 [klein] 4B followed the prompt accurately, its rendering of materials (especially the glass) feels flatter and less immersive.
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 Kontext [max]
- + Excellent shallow depth of field and bokeh effects.
- + Great adherence to the 'imperfect framing' prompt, giving a more candid feel.
- + High-quality reflections on the wet pavement.
- − The bike chain and crankset geometry are physically incoherent.
- − The rain streaks appear a bit too thick and uniform, bordering on a filter effect.
FLUX.2 [klein] 4B
- + Better capture of the 'motion blur from passing cars' in the background.
- + Realistic skin texture and age-appropriate lighting on the subject.
- + The bicycle structure is more grounded and believable despite some gear issues.
- − The subject is not actively 'repairing' the bike; he just seems to be holding it.
- − The framing is very centered and traditional, missing the 'imperfect framing' request.
Verdict: FLUX.1 Kontext [max] delivers a more cinematic and mood-heavy image with superior depth of field and framing that feels like a professional street photograph. FLUX.2 [klein] 4B handles the background motion blur better, but the interaction between the man and the bike feels less like a 'repair' and more like a pose. FLUX.1 Kontext [max] is preferred for its superior artistic composition and adherence to the specific atmospheric prompts.
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 Kontext [max]
- + Exceptional skin texture with realistic pores and sweat details
- + Outstanding lighting that accurately reflects warm tones off the polished engravings
- + Highly detailed hair texture and natural-looking braids
- − The beads in the hair are less distinct compared to those in Model B
- − The extreme close-up limits the view of the leather straps and underlayers
FLUX.2 [klein] 4B
- + Clearly visible small beads integrated into the braids as requested
- + Excellent rendering of the ornate engraved patterns on the armor plates
- + Strong adherence to the requirement for leather straps and cloth underlayers
- − The facial features look a bit too clean and youthful for a 'battle-worn' character
- − The lighting on the skin feels flatter and less immersive than Model A
Verdict: FLUX.1 Kontext [max] produces a significantly more cinematic and lifelike portrait with superior lighting and texture, though it frames the shot very tightly. FLUX.2 [klein] 4B followed more of the specific object prompts like the small beads and multiple layers of clothing, but the overall image quality and 'battle-worn' atmosphere are much stronger in FLUX.1 Kontext [max].
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 Kontext [max]
- + Excellent grid layout with clear spacing
- + Professional typography that mimics a real menu
- + Consistent food photography style
- − Limited food variety (mostly pizza)
- − Small text is largely gibberish
- − The yellow script font is slightly harder to read than the sans-serif
FLUX.2 [klein] 4B
- + Great variety of food photos including salads and proteins
- + Highly vibrant and modern aesthetic
- + Better implementation of the bold sans-serif font requirement
- − Grid layout is a bit cluttered and inconsistent
- − Text includes some distracting malformed characters/symbols
- − Source light in photos varies significantly between tiles
Verdict: FLUX.1 Kontext [max] produces a more cohesive and professional-looking physical menu layout, though it lacks variety in the food shown. FLUX.2 [klein] 4B offers a more vibrant and modern aesthetic with a better selection of dishes (mains vs. just pizza), but the layout is slightly more chaotic. FLUX.1 Kontext [max] is the preferred choice for a clean, professional design that feels ready for use.
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 Kontext [max]
- + Excellent text rendering with no spelling errors.
- + Includes both a full central burger and exploded orbiting ingredients as requested.
- + Clean, professional ad layout with realistic food textures.
- − The price is not inside a starburst as specified.
- − The 'exploded' effect is a bit cluttered with duplicate bun pieces.
FLUX.2 [klein] 4B
- + Correctly places the price within a starburst element.
- + Strong sense of motion with radial blurring in the background.
- + Vivid, high-contrast colors that pop.
- − Significant spelling errors in the primary title text.
- − The burger is mostly intact rather than 'exploded' with all components suspended separately.
- − Text overlaps in a messy way at the top.
Verdict: FLUX.1 Kontext [max] is the clear winner for its perfect handling of the complex text requirements and professional aesthetic. While FLUX.2 [klein] 4B followed the starburst instruction better, it failed significantly on the primary title's spelling and did not accurately depict the 'exploded' burger components specified in the prompt.
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 Kontext [max]
- + Perfect spelling on all requested menu items and the title.
- + Excellent chalk texture and natural variations in the wood frame and board smudges.
- + Clean, readable layout that maintains a consistent handwriting style throughout.
- − The title is in print-style handwriting rather than the requested 'elegant cursive' style.
- − The handwriting looks a bit too 'perfectly' centered, losing a touch of the requested casual slant.
FLUX.2 [klein] 4B
- + Attempts the cursive style for both the title and the body items as requested.
- + Realistic chalk dust and smearing effects across the board surface.
- − Numerous spelling errors including 'Trufffl', 'Musheram', 'Ootrpous', 'Brawn', 'daly', and 'gulten'.
- − The spacing in the top line 'S PECIALS' is broken and awkward.
- − The overall handwriting is less legible compared to Model A.
Verdict: FLUX.1 Kontext [max] is the clear winner due to its flawless text rendering and adherence to complex menu items, despite missing the specific 'cursive' stylistic request for the title. FLUX.2 [klein] 4B followed the stylistic direction for cursive better but failed significantly on basic spelling and character coherence, making the menu largely unreadable.
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 Kontext [max]
- + Excellent anatomical rendering of the horse and suit.
- + High-quality cinematic lighting with realistic lens flare.
- + Clean composition with a clear focal point against the starfield.
- − Completely failed the semantic challenge by placing the astronaut on top instead of the horse.
FLUX.2 [klein] 4B
- + Detailed textures on the astronaut's face and equipment.
- + Good use of perspective with the planet surface below.
- − Failed the semantic challenge by placing the astronaut on top despite the explicit instruction.
- − Anatomical issues with the horse's front legs and movement.
Verdict: Both models failed the specific reasoning component of the prompt, which requested a 'horse riding astronaut' (horse on top). While both FLUX.1 Kontext [max] and FLUX.2 [klein] 4B followed the standard 'astronaut riding a horse' trope, FLUX.1 Kontext [max] delivered a much more polished, cinematic, and anatomically correct image.
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 Kontext [max]
- + Excellent fur texture and lighting on the capybara.
- + Detailed taxi driver hat with clear text.
- + High-quality bokeh effect for the city background.
- − The passenger is holding a phone to her ear rather than looking at it as requested.
- − The perspective hides the passenger's expression and body language.
FLUX.2 [klein] 4B
- + Perfectly captures the passenger looking at her phone with a bored expression.
- + Includes both paws on the steering wheel clearly as requested.
- + The interior composition captures both the driver and the full backseat clearly.
- − The capybara's paws look slightly human-like or distorted.
- − The logo on the hat is illegible compared to Model A.
Verdict: FLUX.2 [klein] 4B followed the prompt's narrative instructions better, specifically capturing the passenger's bored expression while looking at her phone. FLUX.1 Kontext [max] had superior rendering quality and realistic fur textures, but it missed the specific interaction between the passenger and her phone.
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 Kontext [max]
- + Excellent typography with perfect spelling in all requested areas.
- + High-contrast, moody lighting that fits the gothic aesthetic well.
- + Intricate border details with clearly defined webs and thorns.
- − Repetitive text at the bottom, duplicating the location and time information.
- − The scroll banner is less distinct, blending a bit into the background color.
FLUX.2 [klein] 4B
- + Beautiful parchment texture and color palette that feels very 'vintage'.
- + Clear scroll banner element that matches the prompt's layout request.
- + Strong composition with twisted trees framing the central jack-o-lantern.
- − Significant spelling errors in the main title text (e.g., 'Hallbwom').
- − Typographical errors in the date and location labels (e.g., '206' instead of '2026').
Verdict: FLUX.1 Kontext [max] is the clear winner due to its ability to render all requested text accurately and legibly, which is critical for an invitation. While FLUX.2 [klein] 4B provides a lovely vintage atmosphere and parchment effect, the garbled spelling in the title and the missing digit in the date make it unusable as a graphic design piece.
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
FLUX.1 Kontext [max]
- + Successfully added a thick head of hair that matches the facial hair color
- + Preserved most of the background and clothing elements
- − Significantly altered the facial features and glasses of the subject
- − The hairline looks slightly artificial and sits too high on the forehead
FLUX.2 [klein] 4B
- + Excellent preservation of the subject's original facial features, glasses, and expression
- + The hair texture and style look extremely natural and integrate well with the existing beard and lighting
- − Minor blurring where the new hair meets some background branches
Verdict: FLUX.2 [klein] 4B is the clear winner as it successfully adds a realistic head of hair while perfectly preserving the identity of the person in the source image. FLUX.1 Kontext [max] fails the core requirement of preservation by changing the man's face and glasses into a different person entirely.
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 Kontext [max]
- + Excellent typography with clean, consistent 'JAPAN SUSHI' text
- + High visual quality with realistic PBR textures on the fish and rice grains
- + Better adherence to the 'diorama base' and 'isometric miniature' style
- − Missed the small flag icon request
FLUX.2 [klein] 4B
- + Includes a flag icon as requested
- + Clean, high-clarity render with soft lighting
- − Spelling error in the text ('SUSH' instead of 'SUSHI')
- − The flag icon is incorrect for Japan (resembles Austria or Latvia)
- − Less detailed textures on the fish compared to the other model
Verdict: FLUX.1 Kontext [max] is the winner because it successfully rendered the complex text without errors and provided a superior 3D miniature aesthetic with high-quality PBR textures. While FLUX.2 [klein] 4B attempted the flag icon, the spelling error 'SUSH' and the visually flat materials make it a weaker output overall.
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.1 Kontext [max]
- + Expertly captures all three requested themes: TV news, dogs, and hockey.
- + Preserves the denim shirt from the source image while translating it to a caricature style.
- + High-quality, consistent digital illustration style.
- − The hockey element (the stick at top) is a bit disconnected from the main composition.
FLUX.2 [klein] 4B
- + Excellent caricature of the facial features from the source image.
- + Richly detailed TV studio environment that looks professional and high-budget.
- + Includes multiple dogs as requested.
- − Completely misses the hockey requirement.
- − The face has slightly unsettling asymmetric eyes.
Verdict: FLUX.1 Kontext [max] is the winner because it successfully incorporated every element of the prompt, including the hockey stick and the subject's specific denim shirt, into a cohesive caricature. While FLUX.2 [klein] 4B created a more detailed studio environment and did a great job with the facial likeness, it failed to include the hockey theme and had some minor glitches in the character's eyes.
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 Kontext [max]
- + Excellent adherence to the full subject list including the baby bunny.
- + Beautiful soft lighting and glowing god rays.
- + Uniformly high detail on fur and expressive eyes across all animals.
- − The animals are sitting relatively still rather than 'playfully chasing' or 'tumbling'.
- − The butterflies appear somewhat flat and glowing compared to the realistic animals.
FLUX.2 [klein] 4B
- + Successfully captures the action of 'playfully chasing' and 'tumbling'.
- + Great variety in butterfly designs and realistic lighting on the grass.
- + Dynamic composition with a strong sense of movement.
- − Failed to include the baby bunny requested in the prompt.
- − Included two kittens instead of one kitten and one bunny.
- − The lighting is slightly harsh on the fox's face.
Verdict: FLUX.1 Kontext [max] followed the prompt more accurately by including all four specific animals, whereas FLUX.2 [klein] 4B missed the baby bunny entirely. However, FLUX.2 [klein] 4B captured the requested 'playful' and 'tumbling' movement much better than the static pose of FLUX.1 Kontext [max].
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.1 Kontext [max]
- + Perfectly preserves the iconic facial expressions and character poses from the source meme.
- + Accurately captures the specific Ghibli-style character linework and watercolor-esque shading.
- + Maintains the original compositions and background architecture with high fidelity.
- − The color palette is slightly more saturated than the requested 'soft pastel' colors.
FLUX.2 [klein] 4B
- + Executes the 'soft pastel colors' and 'dreamy background' request effectively with a light, airy aesthetic.
- + Good hand-painted texture quality throughout the image.
- − Fails to preserve the critical facial expressions, changing the woman on the right from angry to smiling/neutral.
- − The character on the right appears to be holding a different hand, losing the original context of the 'distracted boyfriend' meme.
Verdict: FLUX.1 Kontext [max] is the superior edit because it successfully translates the image into the Ghibli style while preserving the essential storytelling elements—specifically the characters' facial expressions—that make the source image a recognizable meme. FLUX.2 [klein] 4B provides a beautiful 'dreamy' atmosphere but fails the edit task by changing the expressions and losing the narrative tension of the original photo.
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
FLUX.1 Kontext [max]
- + Successfully added wind-blown hair effect
- + Preserved the original pose and character identity well
- + Added subtle motion with flying leaves
- − The leash and the woman's left hand have become mangled and detached from her body
- − Few leaves added, lacking a truly 'energetic' feel
FLUX.2 [klein] 4B
- + Excellent addition of many dynamic, fluttering leaves
- + Much stronger wind effect in the hair while maintaining facial structure
- + The character's body and the leash remain perfectly intact
- − Some leaves are slightly blurry or lack high-resolution detail
Verdict: While both models successfully added wind effects and flying leaves, FLUX.1 Kontext [max] suffered from significant anatomical errors, detaching the woman's arm and mangling the dog leash. FLUX.2 [klein] 4B provided a much more energetic result with a higher volume of leaves and better hair motion while perfectly preserving the source image's structure.
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 Kontext [max]
- + Perfect text rendering of 'Caffè Florian' and 'Est. 1720'
- + Excellent use of texture to create an authentic vintage woodblock or letterpress feel
- + Superior composition with a beautifully integrated banner and cloche
- − The accent on the 'E' in Caffè is slightly unusual in shape
FLUX.2 [klein] 4B
- + Successfully captured the requested color palette and theme
- + Good cloche illustration with clean vector styling
- − Misspelled the primary name as 'FLAXTION' instead of 'Florian'
- − Redundant text with 'Est. 1720' appearing twice in stacked layers
- − The banner shape is disjointed and awkward
Verdict: FLUX.1 Kontext [max] delivered a professional-grade logo with perfect spelling and a sophisticated vintage texture that feels authentic to the 1720s era. FLUX.2 [klein] 4B failed significantly on the text rendering, misspelling the main name and repeating the date inconsistently.
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 Kontext [max]
- + Excellent text rendering and legibility for major labels.
- + Strong visual storytelling with creative astronaut silhouettes.
- + Clean, professional flat-vector aesthetic that matches the requested NASA palette.
- − Missed several specific steps in the requested sequence (Launch, Earth Orbit).
- − The rocket design is a generic cartoon rocket rather than a Saturn V.
- − Some garbled text in the center and a strange geometric shape above Lunar Orbit.
FLUX.2 [klein] 4B
- + Followed the specific step-by-step instructions and iconography more closely than Model A.
- + Attempted a more accurate representation of the Saturn V rocket structure.
- + Good use of the requested color palette across all elements.
- − Significant spelling and typography errors throughout the poster.
- − Visual layout is cluttered and icons are sometimes overlapping or poorly aligned.
- − The astronaut icons are inconsistent and low quality compared to the rest of the graphic.
Verdict: FLUX.1 Kontext [max] produced a much higher quality, aesthetically pleasing poster with near-perfect title text, but it failed to follow the specific six-step logical flow requested. FLUX.2 [klein] 4B attempted all the specific steps and iconography but suffered from severe text corruption and a messy layout. FLUX.1 Kontext [max] is the preferred choice as its output is usable for a professional context, despite the missing steps.
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