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
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.1 [schnell] FP8
#47 of 62 in Text-to-Image
FLUX.2 [klein] 9B
#13 of 32 in Image Editing
Where the votes landed
FLUX.1 [schnell] FP8
100.0%
win rate
Ties
0.0%
FLUX.2 [klein] 9B
0.0%
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 [schnell] FP8
- + Elegant, sharp geometric lines for the glass cube
- + Vibrant colors and high-end photographic aesthetic
- + Accurate soft lighting from the window side
- − The glass cube contains an illogical horizontal divider/shelf
- − The blue sphere has an Earth-like texture not specifically requested
FLUX.2 [klein] 9B
- + More realistic texture for the wooden table
- + Better physical logic with the sphere sitting at the base of the cube
- + Realistic glass thickness and refraction
- − The sphere appears to be floating slightly above the base
- − The lighting is a bit flat compared to Model A
Verdict: Both models followed the prompt instructions perfectly regarding object placement. FLUX.1 [schnell] FP8 produced a more visually striking, 'clean' image, but it added an unnecessary internal shelf within the cube. FLUX.2 [klein] 9B captured a more grounded, realistic scene with a more convincing wooden texture and simple blue sphere.
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 [schnell] FP8
- + Excellent shallow depth of field and bokeh effect
- + Cinematic color grading with vibrant red reflections
- − The man appears to be simply leaning on the bike rather than repairing it
- − Missing more overt rain textures and tools to support the prompt narrative
FLUX.2 [klein] 9B
- + Stronger narrative adherence with visible tools and a crouching 'repair' pose
- + Realistic skin textures and clothing details with subtle water droplets
- + Includes visible rain and more convincing Japanese street context
- − The bike's front wheel and fork geometry is slightly warped/distorted
- − Less motion blur on cars than requested in the prompt
Verdict: FLUX.2 [klein] 9B followed the narrative aspects of the prompt more effectively by showing the man actually repairing the bike with tools, whereas FLUX.1 [schnell] FP8 felt more like a posed portrait. FLUX.2 [klein] 9B also captured more realistic skin textures and the atmosphere of rain, making it the superior choice for this specific challenge.
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 [schnell] FP8
- + Intense, realistic lighting and shadows
- + Highly detailed facial skin texture
- + Cinematic extreme close-up composition
- − Missed the request for multiple braided hair segments
- − Armor details are partially obscured by the tight framing
FLUX.2 [klein] 9B
- + Excellent adherence to the 'hair braided with beads' prompt
- + High-quality engraving on the plate armor
- + Clear depiction of leather straps over a chainmail/cloth underlayer
- − Slightly less 'gritty' skin texture compared to the other model
- − The background torches feel a bit generic
Verdict: FLUX.2 [klein] 9B followed the specific character instructions much better, capturing the braided hair, beads, and detailed armor layers accurately. While FLUX.1 [schnell] FP8 produced a more intense, cinematically lit portrait with superior skin micro-details, it missed several key descriptive elements 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.”
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent internal grid layout using a white background
- + Includes categorized sections for appetizers, pizza, and mains as requested
- + Clean and professional aesthetic suitable for a real restaurant menu
- − Contains significant spelling errors in headers like 'APPTIZERS', 'PIZZAL', and 'SECCER'
- − The 'MENU' header is cut off at the top of the frame
FLUX.2 [klein] 9B
- + Strong typography and bold sans-serif fonts
- + vibrant accents using gradient underlines and high-saturation photos
- + Clearer text rendering for the primary headers
- − Poor logical mapping, as three distinct category headers ('Appetizers', 'Pizza', 'Mains') all display photos of pizza
- − The main title 'Resteurantbeteina' is nonsensical and has a typo
- − Messy layout in the bottom half with overlapping and inconsistent pricing fonts
Verdict: FLUX.1 [schnell] FP8 is the winner because it creates a more logical and realistic menu layout with distinct photos for different food categories, even though it suffers from significant spelling errors. FLUX.2 [klein] 9B has better color accents and sharper typography, but its failure to match specific food photos to the corresponding headers (showing pizza for every category) makes it less effective as a design mockup.
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 [schnell] FP8
- + Realistic texture on the meat patty and sesame bun
- + Includes various flying ingredient shards for dynamic movement
- − Several spelling errors including 'LIIMITED' and 'NEEY'
- − Failed to use the requested fiery, glowing text effect for 'Magic Burger'
- − Incorrect price display with '€69' taking prominence and '€.99' in the starburst
FLUX.2 [klein] 9B
- + Excellent adherence to the fiery, glowing text effect request
- + Perfect spelling and pricing integration
- + Strong composition with effective use of light trails and embers
- − The burger is less 'exploded' than Model A, appearing more like a standard stacked burger
- − The starburst overlaps the burger bun slightly
Verdict: FLUX.2 [klein] 9B is the clear winner as it followed all complex text instructions perfectly, including the specific fiery glow and correct pricing, whereas FLUX.1 [schnell] FP8 struggled with severe typos and formatting errors. While FLUX.1 [schnell] FP8 had a slightly more 'exploded' look for the burger, the overall visual polish and prompt adherence of FLUX.2 [klein] 9B make it far superior for an advertisement.
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 [schnell] FP8
- + Successfully renders the date correctly.
- + Captures a realistic wooden frame and café lighting.
- − Numerous spelling errors including 'Risortto', 'Griffled', and 'cokies'.
- − Text is repetitive and nonsensical in the lower half.
- − The handwriting looks more like a digital marker than realistic chalk.
FLUX.2 [klein] 9B
- + Excellent adherence to the 'elegant cursive' and 'chalk texture' requirements.
- + Perfect spelling for all requested menu items and prices.
- + Realistic chalk smudge effects and authentic handwriting variations.
- − Slight spelling error in the footer text ('fress' instead of 'fresh').
Verdict: FLUX.2 [klein] 9B followed the prompt instructions near-perfectly, delivering a beautiful chalkboard aesthetic with realistic texture and accurate spelling for the complex menu items. FLUX.1 [schnell] FP8 struggled significantly with the text, producing several overlapping lines of garbled words and failed to capture the requested cursive style.
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 [schnell] FP8
- + Follows the specific prompt instruction to place the horse on top of the astronaut/equipment.
- + Cinematic lighting with a dramatic earth background.
- + Unique interpretation of the astronaut as mechanical equipment being ridden by a horse.
- − Anatomy is confused, resulting in a strange fusion of two horses and a backpack.
- − The astronaut character is almost entirely missing, replaced by a grey robotic box.
FLUX.2 [klein] 9B
- + High visual clarity and vibrant colors in the space background.
- + Well-rendered horse and astronaut details.
- − Completely failed the negative constraint to have the horse on top instead of the human.
- − Includes distracting nonsensical text at the bottom of the frame.
- − The composition is cluttered with too many floating planets.
Verdict: This was a test of prompt adherence regarding spatial relationships. FLUX.1 [schnell] FP8 attempted the difficult 'horse on top' instruction, creating a surreal if somewhat anatomically messy scene, whereas FLUX.2 [klein] 9B defaulted to a standard astronaut-riding-a-horse image, ignoring the specific prompt core. FLUX.1 [schnell] FP8 is the winner for following the unique logic of the request.
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 [schnell] FP8
- + Excellent character rendering with realistic fur texture on the capybara.
- + Vibrant lighting and colors that capture a cinematic New York feel.
- + The capybara's expression is very professional and human-like as requested.
- − The woman in the back is holding two phones, which is a logic error.
- − The capybara's hat is more of a toy hat than a realistic driver cap.
- − The capybara only has one paw on the steering wheel.
FLUX.2 [klein] 9B
- + Perfect adherence to the 'both paws on the steering wheel' instruction.
- + The passenger's expression and posture accurately reflect the 'bored and normal' requirement.
- + Better interior composition that shows the relationship between the front and back seats.
- − The text on the cap's badge is nonsensical ('NEW TALA').
- − The capybara's head is disproportionately large compared to its body and the car interior.
- − The lighting is a bit flat compared to the cinematic quality of the competitor.
Verdict: While FLUX.1 [schnell] FP8 offers superior textures and more vibrant lighting, it fails on logic (two phones) and specific prompt details like the number of paws on the wheel. FLUX.2 [klein] 9B follows the prompt instructions much more closely, capturing the requested composition and Bored expression perfectly, despite some scale issues and garbled text on the hat.
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 [schnell] FP8
- + Features a bold and clear central jack-o-lantern.
- + Captures the moody night sky well with a dark vignette.
- − Significant spelling and grammar errors in all text sections.
- − The border appears more like burnt edges than the requested webs and thorns.
- − Poor typography alignment and layout.
FLUX.2 [klein] 9B
- + Excellent text rendering with no spelling errors.
- + Strict adherence to the requested border of webs and thorns.
- + Higher visual quality with cinematic lighting and a clear gothic aesthetic.
- − The twisted trees are slightly repetitive in form.
- − The moon is partially obscured by the border but still visible.
Verdict: FLUX.2 [klein] 9B followed the prompt instructions perfectly, specifically regarding the complex text requirements and the thematic border of webs and thorns. In contrast, FLUX.1 [schnell] FP8 struggled significantly with the text, producing several typos and failing to include the specific border details requested.
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 [schnell] FP8
- + Excellent soft lighting and refined 3D cartoon textures
- + Clean isometric representation with a professional aesthetic
- + Follows the solid light blue background requirement
- − Text is repeated and misspelled ('JAPAN JAPAN' instead of 'SUSHI')
- − Missing the requested flag icon
FLUX.2 [klein] 9B
- + Perfect text adherence with both 'JAPAN' and 'SUSHI' present
- + Excellent diorama base layout with a wooden block and plate
- + Clear isometric perspective and clean execution
- − Generated the flag of Yemen instead of the Japanese flag
- − Text is a bit close to the top edge of the frame
Verdict: Model B (FLUX.2 [klein] 9B) is the winner because it successfully included all text elements and the diorama base requested in the prompt, whereas Model A failed to include the word 'SUSHI'. While Model B failed on the specific flag nationality, its overall composition and adherence to the layout instructions make it much more useful.
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 [schnell] FP8
- + Features a wider variety of animals and poses.
- + Captures the 'gold rays' and 'dew sparkles' atmospheric effect very well.
- + Excellent soft fur texture and expressive facial features.
- − Failed to include a rabbit, instead adding extra kittens/fox-like creatures.
- − The composition feels slightly crowded with five animals instead of the four requested.
FLUX.2 [klein] 9B
- + More dynamic action with animals actually 'chasing' and 'tumbling' in the scene.
- + Better lighting contrast and sharpness in the wildflower meadow.
- + Higher anatomical accuracy for the specific animals depicted.
- − Completely missed the baby bunny and the red fox kit is the only clearly distinct non-domestic animal.
- − The kitten's facial expression looks somewhat startled rather than playful.
Verdict: Both models struggled to include all four requested animals, specifically failing to generate the baby bunny. FLUX.1 [schnell] FP8 created a more dreamlike, soft-focus masterpiece with beautiful lighting, while FLUX.2 [klein] 9B provided a clearer, more active interpretation of the 'chasing' prompt with better grounding in the meadow. FLUX.2 [klein] 9B is slightly preferred for its superior clarity and more natural interaction between the animals.
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 [schnell] FP8
- + Successfully captures a vintage, ornate aesthetic
- + Includes the established date as requested
- − Significant spelling errors in the main brand name
- − The dome looks more like a building cupola than a food cloche
- − Messy typography arrangement on the banner
FLUX.2 [klein] 9B
- + Perfect text rendering with correct spelling and accents
- + Clear and accurate representation of a food cloche with steam
- + Excellent vector emblem composition and balanced layout
- − The paper texture is slightly more pronounced than 'subtle'
Verdict: FLUX.2 [klein] 9B followed the prompt perfectly, delivering an accurate cloche illustration and flawlessly rendered text. FLUX.1 [schnell] FP8 failed on the text rendering, misspelling the brand name and producing a confusing illustration that looks more like a building than a restaurant cover.
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 [schnell] FP8
- + Successfully adheres to the clean, minimalist flat-vector style requested.
- + Follows the NASA-inspired color palette accurately with a professional navy background.
- + Consistent and organized layout that feels like a real infographic.
- − Text is largely gibberish or contains significant typos throughout the image.
- − Icons are very abstract to the point of being unidentifiable without the text labels.
FLUX.2 [klein] 9B
- + Illustrations are clear and much more representative of the subjects (Saturn V, Lunar Module).
- + Typography is more legible, including a successful rendering of 'Tranquility' and 'Armstrong'.
- + Visual storytelling is more engaging with varied icon sizes and compositions.
- − Failed the color palette prompt by including bright greens and oranges.
- − The layout is a bit cluttered compared to the 'clean' request.
- − Spelling errors in main headers like 'AOILLO' and 'LANNDINING'.
Verdict: FLUX.1 [schnell] FP8 better captures the 'clean vector' aesthetic and strict color palette requested in the prompt, despite the illegible text. FLUX.2 [klein] 9B provides much better illustrative detail and clearer iconography, but misses the mark on the specific NASA color constraints and has a slightly messy arrangement.
Explore each model
Black Forest Labs' distilled 9 billion parameter image generation model with sub-second inference and multi-reference support