Black Forest Labs' precision image generation model with maximum control, reliable text rendering, and complete creative control supporting up to 4MP output
Settled by community votes across 19 shared challenges, with an AI judge weighing in on each.
FLUX.2 [flex]
#14 of 62 in Text-to-Image
FLUX.2 [klein] 4B
#32 of 62 in Text-to-Image
Where the votes landed
FLUX.2 [flex]
0.0%
win rate
Ties
0.0%
FLUX.2 [klein] 4B
100.0%
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.2 [flex]
- + Excellent adherence to the 'small blue sphere' prompt requirement
- + Highly realistic rendering of the wooden table and book textures
- + Precise glass reflections and refractions of the background plant
- − The sphere is slightly large for the description 'small' compared to the cube interior
FLUX.2 [klein] 4B
- + Better interpretation of the 'small' scale for the blue sphere
- + Very natural lighting and depth of field
- + Realistic imperfections on the glass surface
- − The red book appears significantly thinner and less substantial than expected
- − Minor blurring on the edges of the book binding
Verdict: Both models followed the complex spatial instructions perfectly. FLUX.2 [flex] is preferred for its superior texture rendering and sharp clarity, while FLUX.2 [klein] 4B handled the relative scale of the 'small' sphere slightly better but had weaker rendering on the book.
Man and Car in California
Editing“Make a photo of the man driving the car down the California coastline”
AI Judge Analysis
FLUX.2 [flex]
- + Excellent preservation of the car's specific make and model details.
- + Accurately places the man from the source image behind the wheel.
- + Good sense of motion through wheel blur and road lines.
- − The man's outfit was changed from the source image to a plain shirt.
FLUX.2 [klein] 4B
- + Successfully carries over the patterned coat from the source image for the driver.
- + Highly realistic California coastline environment.
- + Preserves the car's design elements accurately.
- − The driver's facial features and hair lose some resemblance to the source image compared to the other model.
- − The car's headlights appear slightly distorted.
Verdict: Both models performed excellently at this complex image editing task, correctly merging two subjects into a completely new environment. FLUX.2 [flex] achieved a better facial likeness for the man, while FLUX.2 [klein] 4B did a superior job of preserving the man's specific clothing from the source image.
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 [flex]
- + Excellent shallow depth of field and bokeh with 50mm lens characteristic.
- + Superior rendering of wet pavement and light reflections.
- + The bicycle mechanics and the man's interaction with the chain are realistic.
- − Stronger vertical crop makes it feel less like a candid wide-street shot.
FLUX.2 [klein] 4B
- + The wider framing captures the 'candid street' atmosphere better.
- + Good use of motion blur on the background vehicles.
- + Natural rain textures on the subject's clothing.
- − The man is oddly sitting on/through the bicycle frame rather than repairing it.
- − Anatomical issues with the man's hands and the bicycle's geometry.
- − The man's right leg appears to be missing or merged into the bike frame.
Verdict: FLUX.2 [flex] produced a much more coherent and realistic image, particularly in the mechanical details of the bicycle and the man's posture. While FLUX.2 [klein] 4B followed the 'street photo' composition well, it failed significantly on anatomical and object coherence, making the man appear to be fused with the bicycle.
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 [flex]
- + Excellent rendition of the engraved plate armor with high-contrast detail.
- + Superior lighting with warm torchlight glints and strong metallic reflections.
- + Highly detailed texture on the leather straps and knitted cloth underlayer.
- − The sparks in the background are somewhat uniform and circular.
FLUX.2 [klein] 4B
- + Clean facial features with realistic skin texture and subtle scars.
- + Good adherence to the braided hair and beads request.
- + Effective use of shallow depth of field for the background.
- − The armor engraving is less intricate and flatter compared to the other model.
- − The leather straps lack the requested high-detail texture.
- − Lighting on the armor feels slightly matte rather than reflecting the warm torchlight vividly.
Verdict: FLUX.2 [flex] is the clear winner for its superior texture work, especially on the engraved armor and the leather straps, which feel much more lifelike. While FLUX.2 [klein] 4B provides a clean portrait, it fails to capture the rich material interplay and warm, dramatic lighting that FLUX.2 [flex] achieves.
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 [flex]
- + Excellent typographic layout with clearly defined headers for Pizza and Mains.
- + The food photography is high-quality and directly relates to the category headers.
- + Highly readable fonts and consistent price formatting.
- − The main header incorrectly says 'APPETIZERS' for the entire page despite more specific sub-headers.
- − Text descriptions below menu items are garbled gibberish.
FLUX.2 [klein] 4B
- + Strong grid-based layout for photography that feels modern and artistic.
- + Distinct branding and a more 'vibrant' accent color palette as requested.
- − Typos in major section headers like 'MDAINS' and 'FRAZZA'.
- − The prices and fonts are inconsistent and less legible than the other model.
- − The pizza photo is placed far away from the relevant text section.
Verdict: FLUX.2 [flex] produced a vastly superior layout that looks like a real-world usable menu, featuring clean typography and food images that actually match their labels. FLUX.2 [klein] 4B struggled with spelling in the headers and created a disjointed relationship between the food photos and the text sections.
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.2 [flex]
- + Perfect text rendering for all requested phrases including the price.
- + Excellent 'exploded' effect with distinct suspended layers and debris.
- + High-quality textures on the meat patty and fresh vegetables.
- − The bottom lettuce layer is slightly repetitive in shape.
FLUX.2 [klein] 4B
- + Good sense of motion with radial blurring in the background embers.
- + Accurate rendering of the price in the starburst element.
- − Severe text corruption in the primary title rendering it illegible.
- − Failed to create an 'exploded' burger, showing a mostly assembled product.
- − The burger components lack the photorealistic crispness seen in the competitor.
Verdict: FLUX.2 [flex] successfully followed every instruction, providing a high-quality exploded burger view and perfectly rendered, fiery text. FLUX.2 [klein] 4B failed significantly on the text rendering of 'MAGIC BURGER' and did not achieve the requested 'exploded' composition, leaving the burger largely intact.
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.2 [flex]
- + Perfect spelling of all menu items and dates.
- + Excellent chalkboard texture with realistic smudges and chalk dust.
- + Consistent and elegant cursive handwriting style that looks authentic.
- − The font is slightly too clean, appearing almost like a digital chalk-style font rather than purely manual handwriting.
FLUX.2 [klein] 4B
- + Great chalk texture and stroke variation that feels very 'handwritten'.
- + Good use of the chalkboard space and frame details.
- − Numerous spelling errors including 'Trufffl', 'Musheram', 'Ootrpous', 'Brawn', and 'daly'.
- − Significant spacing issues in the header text 'SPECIALS'.
Verdict: FLUX.2 [flex] is the clear winner as it followed the complex text prompt perfectly with no spelling errors, whereas FLUX.2 [klein] 4B struggled significantly with spelling and character spacing. FLUX.2 [flex] also managed to create a much more convincing atmosphere and professional layout for a café setting.
Pose & Character Mashup
Editing“Use Image 1 as the exact pose reference and Image 2 as the character reference. Recreate the person/character from Image 2 in the exact dynamic pose and body position from Image 1. Keep the exact face, hair, clothing style/details, and expression from Image 2. Match the lighting and environment of Image 1. The final image must show the character from Image 2 performing the precise action/pose from Image 1 with perfect anatomy and natural integration.”
AI Judge Analysis
FLUX.2 [flex]
- + Excellent character facial recognition and clothing detail preservation
- + Matches the lighting and background of the source image perfectly
- + Maintains clear, high-resolution details on the scarf and outfit
- − Fails to replicate the specific crossed-leg, leaning pose from Image 1
- − Anatomical issues with the raised arm and hand positioning
FLUX.2 [klein] 4B
- + Successfully replicates the exact complex pose from Image 1
- + Correctly blends character features like the sunglasses and clothing style onto the dynamic pose
- + Maintains the long hair texture from the original pose image for better integration
- − The facial likeness is slightly distorted due to the angle
- − The scarf and shoulder area appear messy and poorly defined
Verdict: FLUX.2 [klein] 4B is the clear winner as it successfully follows the most difficult instruction: replicating the exact, complex pose of the person on the stool. While FLUX.2 [flex] has better overall image quality and facial likeness, it completely ignores the body position and lean of Image 1, producing a standard crouching pose instead.
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.2 [flex]
- + Excellent prompt adherence by placing the horse on top of the astronaut.
- + High creative merit with a surreal, cinematic composition.
- + Rich colors and detailed textures on both the horse and the space suit.
- − Anatomical issues where the horse's front legs merge oddly with the astronaut's shoulders.
FLUX.2 [klein] 4B
- + High visual clarity and realistic lighting on the astronaut's suit.
- + Good representation of the space environment with the Earth below.
- − Failed the negative constraint by putting the astronaut on top of the horse.
- − Composition is quite conventional and lacks the requested surrealism.
Verdict: FLUX.2 [flex] successfully followed the difficult prompt instruction to place the horse on top of the astronaut, resulting in a unique and surreal image. FLUX.2 [klein] 4B ignored the specific spatial constraint, producing a standard astronaut-riding-horse image which failed the primary challenge of the prompt.
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.2 [flex]
- + Excellent photorealistic lighting and skin/fur textures
- + Highly detailed taxi interior that looks authentic
- + Superior composition with a cinematic wide-angle feel
- − The capybara's hands look more human-like than paws
- − The passenger is partially obscured by the headrest
FLUX.2 [klein] 4B
- + Successfully shows both front paws on the steering wheel
- + Accurately captures the specified 'bored' expression on the businesswoman
- − Poor spatial logic/composition; the passenger appears to be sitting next to the driver in a tiny car rather than in the back seat
- − Lower overall image resolution and clarity compared to Model A
- − The capybara's cap looks like a modern baseball cap rather than a traditional driver cap
Verdict: FLUX.2 [flex] produced a much higher quality, cinematic image with superior lighting and texture detail, though it struggled with the anatomical accuracy of the paws. FLUX.2 [klein] 4B followed the prompt instructions for the paws and expression better, but failed significantly on the composition, creating a cramped and unrealistic vehicle interior that lacked the depth and professional finish of the first model.
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.2 [flex]
- + Perfect text rendering for all requested strings.
- + Excellent cinematic lighting on the jack-o-lantern.
- + Strong composition with a clear, readable layout.
- − The parchment border effect is slightly generic.
FLUX.2 [klein] 4B
- + Atmospheric integration of twisted trees and foggy background.
- + The web and thorn border is very detailed and fits the prompt well.
- − Significant spelling errors in the title and event details.
- − The year is missing a digit (206 instead of 2026).
- − The scroll banner text has a typo ('friglts').
Verdict: FLUX.2 [flex] successfully rendered all text elements accurately, which is crucial for a party invitation. In contrast, FLUX.2 [klein] 4B struggled significantly with spelling, resulting in garbled text in the title and missing information in the date. FLUX.2 [flex] is the clear winner for its polish and prompt adherence.
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.2 [flex]
- + Excellent preservation of the original image's lighting, facial features, and background.
- + Realistic hair texture and color that matches the existing beard.
- − The hairline is a bit high and stiff, feeling more like a buzz cut than a full head of hair.
FLUX.2 [klein] 4B
- + Very creative and thick hair style that fits the requested description well.
- + Natural-looking waves and volume that blend into the original facial hair.
- − Slightly alters the shape of the brow and forehead compared to the source image.
- − Visible artifacting where the hair meets the background sky.
Verdict: FLUX.2 [flex] provides a more conservative edit that perfectly preserves the original subject's details, though its hair addition is relatively short. FLUX.2 [klein] 4B followed the instruction for 'full, thick' hair more literally with a longer style, but it introduced slight changes to the facial structure and some blending artifacts against the sky. FLUX.2 [flex] is the winner for its superior source preservation and integration.
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 [flex]
- + Perfect text rendering for both 'JAPAN' and 'SUSHI'.
- + Accurately includes the Japanese flag as requested.
- + Excellent miniature 3D aesthetic with a clean diorama base.
- − The salmon texture is slightly over-simplified compared to real sushi.
FLUX.2 [klein] 4B
- + The sushi fish texture has high realism and great PBR material qualities.
- + Good adherence to the 45-degree isometric angle.
- − Text is misspelled as 'SUSH'.
- − The flag icon is incorrect, appearing more like the flag of Austria or Latvia.
- − Lacks the 'diorama base' requested, featuring only a plate.
Verdict: FLUX.2 [flex] followed every instruction perfectly, including the specific text, the correct national flag, and the diorama base. FLUX.2 [klein] 4B failed on the text spelling and the flag icon, though it did produce a more realistic material texture for the salmon.
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 [flex]
- + Perfectly incorporates all three elements: TV anchor, dogs, and hockey jerseys/gear.
- + Maintains the subject's denim shirt style from the original image while translating it to a caricature.
- + Highly creative and cohesive composition with a clear newsroom theme.
FLUX.2 [klein] 4B
- + Strong caricature style with exaggerated facial features.
- + Good rendering of the news anchor desk environment and two dogs.
- − Completely misses the 'hockey' requirement from the prompt.
- − Leaves the source image's identity behind, making the subject look like a different person.
Verdict: FLUX.2 [flex] is the clear winner as it successfully incorporated every element of the prompt, including the hockey theme which FLUX.2 [klein] 4B overlooked. Additionally, FLUX.2 [flex] managed to preserve the subject's clothing style (denim shirt) from the source image, making the caricature feel more like a direct transformation.
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 [flex]
- + Successfully included all four requested animals (dog, cat, bunny, fox).
- + Excellent rendering of 'god rays' and dew sparkles in the grass.
- + Clearer 8K resolution with authentic fur textures and realistic eyes.
- − The fox's anatomy is slightly elongated in the body.
FLUX.2 [klein] 4B
- + Captured a high sense of playfulness and tumbling action between the animals.
- + Beautiful lighting and vibrant color palette in the flowers.
- − Failed to include the requested baby bunny.
- − Included two kittens instead of one kitten and one bunny.
- − The fox's facial features look slightly more illustrative/simplified compared to the dog.
Verdict: FLUX.2 [flex] is the clear winner as it strictly followed the prompt by including all four distinct animal types, whereas FLUX.2 [klein] 4B missed the baby bunny entirely. FLUX.2 [flex] also achieved a more hyper-photorealistic look with superior texture detail in the fur and environment.
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 [flex]
- + Excellently captures the specific Studio Ghibli art style with flat cel-shading and characteristic facial features.
- + Maintains the exact essence and composition of the source meme perfectly.
- + The lighting and background textures feel authentic to hand-painted anime backgrounds.
- − The character in blue loses the 'angry/shocked' expression from the original meme, looking too neutral/happy.
FLUX.2 [klein] 4B
- + Beautiful use of soft pastel colors and glow effects.
- + Highly detailed line work and lovely coloring on the clothing.
- + Good representation of the nostalgic, dreamy mood requested.
- − Loses the specific Ghibli facial aesthetic, leaning more towards modern shoujo manga.
- − Significantly alters the background from the source image into a garden-like setting.
- − The girlfriend's expression is smiling, which completely misses the context of the original meme.
Verdict: FLUX.2 [flex] wins because it successfully adheres to the specific Ghibli art style while preserving the narrative context and composition of the original 'Distracted Boyfriend' meme. While FLUX.2 [klein] 4B produces a very pretty image with great colors, it fails to capture the iconic Ghibli character design and incorrectly changes the jealous girlfriend's expression to a friendly one.
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.2 [flex]
- + Excellent preservation of the subject's face and original clothing details
- + Naturalistic wind-blown hair effect
- − The flying leaves look somewhat repetitive and like simple overlays
- − The dog's tail motion remains static despite requested 'dynamic motion'
FLUX.2 [klein] 4B
- + High density and variety of flying leaves creates a strong sense of wind
- + Good hair movement that feels integrated with the scene
- − The leaves exhibit some motion blur while the hair is oddly sharp, creating a slight visual disconnect
- − The dog's tail appears slightly distorted compared to the original
Verdict: Both models followed the instructions well, successfully adding wind-blown hair and leaves. FLUX.2 [flex] produced a cleaner image with better subject preservation, while FLUX.2 [klein] 4B created a more 'energetic' feel by adding a larger number of variegated leaves across the entire frame.
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 [flex]
- + Excellent typography with perfect spelling and accentuation
- + Clean, professional vector lines suitable for a logo
- + Balanced circular composition
- − Minimalist background texture is barely visible
FLUX.2 [klein] 4B
- + Classic vintage aesthetic with good paper texture
- + Nice use of color blocking on the cloche
- − Significant spelling error in 'FLORIAN' which is rendered as 'FLAXTION'
- − Repetitive information with the date appearing twice in two different styles
- − Inconsistent line weights and messy banner rendering
Verdict: FLUX.2 [flex] produced a professional-grade vector logo that followed all prompt instructions, including the spelling and accents of 'Caffè Florian'. FLUX.2 [klein] 4B failed on basic text accuracy and included redundant elements that cluttered the design.
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 [flex]
- + Excellent typography with perfect spelling and clear hierarchy.
- + Follows the NASA-inspired color palette and flat-vector style accurately.
- + Clean, logical layout that chronologically follows the mission steps.
- − Missing the final sixth step (Landing) as it stops at Descent.
- − The connection between the TLI arc and the rest of the flow is slightly disjointed.
FLUX.2 [klein] 4B
- + Includes a representation of the landing surface and astronaut silhouettes.
- + Attempts to use consistent iconography for the planetary bodies.
- − Extremely poor text rendering with major spelling errors throughout.
- − The Saturn V rocket design is inaccurate and looks more like a modern generic rocket.
- − Cluttered composition with illogical sequencing of the mission steps.
Verdict: FLUX.2 [flex] is the clear winner as it produces a professional, clean infographic with perfect English text and a cohesive design style. FLUX.2 [klein] 4B fails significantly on text legibility and the logical flow of the infographic, resulting in nonsensical labels like 'APLLO 1UT 1J'.
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