Black Forest Labs' 12 billion parameter distilled image generation model optimized for speed, capable of generating high-quality images in just 4 inference steps
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.1 [schnell]
18.7
arena score
#48 of 62 in Text-to-Image
Skill signature
· Text-to-Image
FLUX.1 [schnell] FP8
19.0
arena score
#46 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.1 [schnell]
0%
win rate
Ties
0%
FLUX.1 [schnell] FP8
0%
win rate
Shared challenges
13
Challenge by challenge
The strongest take from each model on every shared challenge, with the AI judge's read.
Geometric Composition
Text-to-Image“A glass cube on a wooden table. Inside the cube is a small blue sphere. On top of the cube sits a red book. A green plant is behind the cube, partially visible through the glass. Soft window light from the left.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent depiction of a monstera plant behind the glass
- + Includes all requested elements correctly
- + Realistic soft lighting from the window side
- − Added an extra blue sphere on top of the book
- − The blue sphere inside the cube is floating without physics
- − The cube edges are slightly distorted near the top
FLUX.1 [schnell] FP8
- + Successfully placed the sphere inside the cube on a surface
- + Excellent glass refraction and transparency effects
- + High visual clarity and realistic wood grain
- − The glass container is more of a tall rectangular prism than a cube
- − The plant is positioned more to the side than directly behind the glass view
Verdict: Both models followed the complex spatial instructions well, but Model B (FLUX.1 [schnell] FP8) is the winner as it adhered more closely to the object count, whereas Model A added an extra sphere. Model B also featured more realistic physics and superior glass rendering, despite the cube being slightly elongated.
Candid Street Photography
Text-to-Image“A candid street photo of an elderly Japanese man repairing a red bicycle in light rain, reflections on wet pavement, shallow depth of field, 50mm lens, natural skin texture, imperfect framing, motion blur from passing cars, cinematic but realistic, no stylization.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent depiction of rainy atmosphere with visible reflections and a cold, damp color grade.
- + Realistic skin texture and age details on the subject.
- + Stronger sense of 'candid' street photography with the imperfect framing and closer perspective.
- − The bike anatomy is slightly flawed with the chain disappearing into a guard that doesn't align well.
- − Failed to include any noticeable motion blur from passing cars.
FLUX.1 [schnell] FP8
- + Natural composition and very high image clarity.
- + Accurate reflections of taillights on the wet pavement.
- + Good sense of depth with professional-looking bokeh.
- − Lacks the requested 'motion blur' for the car in the background which looks static.
- − The subject feels a bit more posed than 'candid' compared to the other model.
- − Bicycle handlebars and brake cables are structurally confusing.
Verdict: Both models captured the essence of the prompt, but FLUX.1 [schnell] (Model A) produced a more convincing atmosphere and texture that felt like a real candid shot in the rain. FLUX.1 [schnell] FP8 (Model B) had cleaner colors but missed the gritty 'realistic/no stylization' feel requested in the prompt. Both models failed to adequately represent the requested motion blur from passing cars.
Fantasy Warrior
Text-to-Image“Close portrait of a battle-worn paladin in ornate engraved plate armor, hair braided with small beads, faint scars and dirt on the skin, warm torchlight reflecting off metal, shallow depth of field, bokeh sparks, lifelike eyes, highly detailed texture on leather straps and cloth underlayer.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent skin texture with visible pores and realistic weathering
- + Intricate engraving details on the armor plates
- + Strong intensity in the gaze with clear eye reflections
- − Hair beads are present but less distinct as individual decorative elements
- − The lighting on the face is slightly flat compared to the armor
FLUX.1 [schnell] FP8
- + Superior use of torchlight lighting, creating dramatic shadows and highlights on the face
- + Better adherence to the 'beads' hair detail with visible metallic spheres
- + Exceptional detail on the cloth underlayer and leather straps with embossed patterns
- − Slightly more artificial 'painted' look to some of the skin transitions
- − Small artifacting on the lower right shoulder with garbled text-like markings
Verdict: Both images are remarkably high quality, appearing to be variations from the same model family. FLUX.1 [schnell] FP8 is the winner because it better captures the dramatic 'warm torchlight' requested and provides more distinct details for the requested hair beads and leather textures, despite a minor artifact on the armor.
Modern Clean Menu
Text-to-Image“Modern minimalist restaurant menu design, white background with colorful food photos in grid, sections for appetizers/pizza/mains, bold sans-serif fonts, vibrant accents, clean professional layout for casual dining.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent typography with clean, legible sans-serif fonts
- + Professional minimalist layout that looks like a real menu design
- + Good sectioning for Appetizers, Pizza, and Mains.
- − The food photos are repetitive, showing multiple pizzas in the appetizers section
- − Small body text is illegible gibberish.
FLUX.1 [schnell] FP8
- + More variety in food photography colors and textures
- + Comprehensive layout including pricing and descriptions
- + Effective use of grid-based design for a two-page spread.
- − Contains typos in major headings like 'Apptizers' and 'Pizzal'
- − The top 'Menu' text is cut off at the page edge
- − Confusing section header 'Seccer' makes no sense in a food context.
Verdict: FLUX.1 [schnell] is the winner because it adheres much better to the minimalist aesthetic and maintains a professional, clean layout without the glaring header typos found in its counterpart. While FLUX.1 [schnell] FP8 offers more visual variety in the food photos, the cut-off 'Menu' text and words like 'Seccer' and 'Apptizers' ruin the professional design requested.
Magic Burger Explosion: Fiery Photorealism Challenge
Text-to-Image“Ad for 'Magic Burger'. Dynamic, exploded burger with all components (bun, patty, cheese, lettuce, tomato, sauce) suspended in mid-air. Emphasize photorealistic detail and a sense of motion. Dark, fiery background with glowing embers. Integrate text: 'MAGIC BURGER' as a prominent title, 'LIMITED TIME ONLY' as a secondary message, and '€6.99' in a starburst, all rendered with a fiery, glowing effect.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + The burger texture is highly photorealistic with appealing melting cheese.
- + The overall composition feels balanced and professional.
- + Captures the fiery background aesthetic well.
- − Failed the primary title text by rendering 'AGIC BURGER' instead of 'MAGIC BURGER'.
- − The price starburst is cluttered with three different versions of the numbers.
- − Individual components like croutons/nuggets feel slightly random compared to a standard burger assembly.
FLUX.1 [schnell] FP8
- + Successfully spelled 'MAGIC BURGER' in the primary title.
- + The 'exploded' effect with mid-air ingredients is more dynamic and varied.
- + Good lighting on the food items.
- − Serious spelling errors in the secondary text including 'LIIMITED' and 'NEEY'.
- − Failed to render the correct price, showing '€69' and '€.99' instead of '€6.99'.
- − Text lacks the requested fiery, glowing effect, appearing as flat white and yellow.
Verdict: FLUX.1 [schnell] creates a much more appetizing and photorealistic burger, but fails the main title spelling and clusters the price information poorly. FLUX.1 [schnell] FP8 gets the main title correct but suffers from severe typos in the subtext and incorrect pricing. FLUX.1 [schnell] is preferred as a base image because its food quality is superior and its single missing letter is easier to edit than the multiple text hallucinations in the other model.
Chalkboard Menu
Text-to-Image“Handwritten-style chalkboard menu in a cozy café, all text rendered in the exact same realistic chalk handwriting style with natural variations in letter size, slight slant, and chalk texture — no printed or digital fonts anywhere on the board. Title at the top in elegant cursive chalk handwriting: ‘TODAY’S SPECIALS – APRIL 30, 2026’. Below it, three menu items also in the same handwritten chalk style: ‘Truffle Mushroom Risotto – $24’, ‘Grilled Octopus with Lemon & Herbs – $28’, ‘Brown Butter Chocolate Chip Cookies – $9’. At the very bottom, smaller text in the identical handwritten chalk style (slightly smaller but still clearly legible with the same handwriting characteristics): ‘All items made fresh daily • Ask about our gluten-free options’. Warm ambient café lighting, visible chalk dust and smudges, realistic handwriting imperfections, no clean printed text anywhere.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Legible chalk-like texture in the handwriting.
- + Maintains a consistent vertical alignment for the menu items.
- − Major spelling errors in the title ('Pril' instead of 'April 30') and most menu items.
- − Handwriting is too uniform and lacks the 'elegant cursive' requested for the title.
FLUX.1 [schnell] FP8
- + Correctly spells out 'APRIL 30, 2026' in the header.
- + Better variety in font weights and sizes, reflecting natural chalk variations.
- − Text layout becomes cluttered and repetitive toward the middle.
- − Fails to render 'elegant cursive' for the title, using a blocky style instead.
Verdict: Both FLUX.1 [schnell] and the FP8 version struggle significantly with the complex instruction for spelling and handwriting style. FLUX.1 [schnell] FP8 is slightly better as it correctly renders the date in the title, whereas the standard model cuts the month in half, though both models failed the request for elegant cursive and accurate menu text.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Successfully placed the horse on top of the astronaut as requested.
- + The lighting on the astronaut's suit is cinematic and matches the planetary background.
- + Features a creative, surreal harness and saddle setup linking the two figures.
- − The horse has two heads growing out of a single neck.
- − The astronaut's anatomy is mangled, with a helmet appearing at the abdomen area.
FLUX.1 [schnell] FP8
- + High resolution with sharp details on the horse's fur.
- + Maintains a clear 'horse on top' composition relative to the spaceship/backpack.
- − Completely missing the astronaut figure, showing only a mechanical pack.
- − Features the same two-headed horse anatomical error as seen in the other model.
- − Unrelated architectural element at the bottom creates visual clutter.
Verdict: Both FLUX.1 [schnell] and FLUX.1 [schnell] FP8 struggled with the surreal anatomical constraints, resulting in a two-headed horse in both instances. FLUX.1 [schnell] followed the prompt more accurately by including an identifiable astronaut, even though the body was distorted, whereas FLUX.1 [schnell] FP8 replaced the human entirely with a mechanical module.
The Capybara Taxi Driver
Text-to-Image“Photorealistic scene inside a yellow New York taxi at night. A capybara is driving, wearing a yellow taxi driver cap and a dark jacket. It has a calm, professional expression and both front paws on the steering wheel. In the back seat sits a human businesswoman in a coat, looking at her phone with a completely normal, bored expression (as if this is just another normal ride). Through the windows you can see the streets of Manhattan at night with blurred lights. Realistic taxi interior, photorealistic, detailed fur and fabric, 35mm lens, night lighting with reflections, shallow depth of field.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent fur texture and lighting on the capybara.
- + Successfully followed the paw placement instruction on the steering wheel.
- + The passenger's expression and posture perfectly match the 'bored' instruction.
- − One of the capybara's eyes appears slightly distorted or glazed.
- − The taxi interior feels a bit flat in the foreground.
FLUX.1 [schnell] FP8
- + Stronger sense of depth with the exterior Manhattan street lights.
- + Very clean rendered text on the taxi cap.
- + Good lighting coherence across the whole scene.
- − The passenger is strangely holding two phones, which was not requested.
- − The capybara's paw placement is less natural compared to Model A.
- − The passenger's expression is slightly more pleasant/smiling than the 'bored' requirement.
Verdict: FLUX.1 [schnell] followed the prompt more accurately, particularly regarding the capybara's interaction with the steering wheel and the passenger's bored expression. FLUX.1 [schnell] FP8 introduced an odd artifact of the passenger holding two smartphones, although its background rendering was slightly more vibrant.
The Halloween Invitation
Text-to-Image“Vintage gothic Halloween party invitation. Dark parchment poster, spooky border with webs and thorns, central glowing jack-o-lantern, bats, twisted trees, moody night sky. Add elegant gothic title text saying "Halloween Party Invitation", a small scroll banner saying "You are invited to a night of frights", and event details at the bottom: Date: 30.10.2026 Time: 7pm Location: The Arches, NYC Spooky but polished, cinematic lighting, square format.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Successfully included almost all specific event details including the correct date and location name.
- + Excellent atmosphere with moody lighting and a clear jack-o-lantern focal point.
- + Good use of the border element with subtle thorn and web details.
- − Has repetitive text lines and accidental gibberish at the bottom ('Fate: butistigtion').
- − Includes multiple banners when only one scroll banner was requested.
FLUX.1 [schnell] FP8
- + High-quality rendering of the jack-o-lantern with a more realistic, detailed texture.
- + Captures the 'twisted trees' and 'spooky border' requirements very effectively.
- + Composition feels slightly more balanced and centered.
- − Significant spelling errors throughout all text sections ('Tine', 'Itrme', 'Theaches').
- − Failed to include the specific date format and full location name correctly.
Verdict: Both models effectively captured the visual aesthetic of a gothic Halloween invitation, but FLUX.1 [schnell] is the superior choice because it managed to render the specific event details (Date and Location) with much higher accuracy. FLUX.1 [schnell] FP8 struggled significantly with text rendering, resulting in numerous typos and missing specific data points requested in the prompt.
Isometric Miniature Diorama Scenes
Text-to-Image“Create a clear, 45° top-down isometric miniature 3D cartoon scene of Japan's signature dish: sushi, with soft refined textures, realistic PBR materials, gentle lighting, on a small raised diorama base with minimal garnish and plate. Solid light blue background. At top-center: 'JAPAN' in large bold text, 'SUSHI' below it, small flag icon. Perfectly centered, ultra-clean, high-clarity, square format.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent text rendering and placement of the word JAPAN.
- + High visual clarity in the miniature diorama aesthetics.
- + Realistic procedural textures on the salmon and rice.
- − Completely missed the word SUSHI below the main title.
- − The red markings on the salmon look a bit like messy sauce rather than natural marbling.
FLUX.1 [schnell] FP8
- + Presents a nice variety of sushi types on the platter.
- + Faithful to the 45-degree isometric projection requested.
- − Failed the text instruction with garbled 'JAPAN' and missing 'SUSHI'.
- − Missing the flag icon, instead trying to integrate the circle into the text.
- − Textures appear slightly more plastic-like compared to the realistic PBR request.
Verdict: FLUX.1 [schnell] is the winner for its superior text rendering and overall image clarity, despite missing the second word requested. FLUX.1 [schnell] FP8 struggled significantly with the typography and diorama composition, resulting in a cluttered and less accurate representation of the prompt.
Adorable Baby Animals in Sunny Meadow
Text-to-Image“Hyper-photorealistic scene of fluffy baby animals—a golden retriever puppy, tabby kitten, baby bunny, and red fox kit—with big expressive eyes and ultra-detailed soft fur, playfully chasing butterflies and tumbling together in a lush wildflower meadow, warm golden sunrise light with god rays and dew sparkles, joyful wholesome vibe, 8K masterpiece.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Excellent fur texture and individual fiber detail.
- + Good spatial arrangement and separation of the animals.
- + Natural lighting that creates a soft, warm atmosphere.
- − One creature appears to be a feline-rabbit hybrid rather than a distinct baby bunny.
- − The animals are mostly stationary rather than in a 'tumbling' or 'chasing' action.
FLUX.1 [schnell] FP8
- + Successfully captures a more playful and dynamic energy with the kittens reaching for butterflies.
- + Beautifully rendered god rays and golden hour lighting.
- + Includes a larger group of animals which feels more like a 'tumbling' scene.
- − Failed to provide a clear baby bunny, instead providing multiple kittens and foxes.
- − Some minor anatomical issues where animals overlap in the center.
- − The lower-right creature has slightly distorted facial features.
Verdict: FLUX.1 [schnell] provides a cleaner, more realistic image with better individual subject clarity, though it struggles with the 'bunny' prompt. FLUX.1 [schnell] FP8 captures the 'playful' and 'god rays' aspect of the prompt more effectively but suffers from more anatomical confusion and also fails to generate a distinct rabbit. FLUX.1 [schnell] is the likely winner for its superior technical polish and realistic fur rendering.
Vintage Cafe Logo
Text-to-Image“Vintage minimalist restaurant logo for "Caffè Florian", retro cloche dome with steam and "Est. 1720" banner, classic typography, warm brown and cream tones, subtle texture on light background, vector emblem style.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Clean vector-style execution with high clarity
- + Consistent line weight and professional spacing
- + Adheres well to the color palette and minimalist style
- − Significant text errors including 'FRAMILAN' and '7720'
- − Missing the requested steam element
FLUX.1 [schnell] FP8
- + Successfully includes the steam and banner elements
- + Correctly renders the year '1720'
- + Elegant cloche design with detailed flourishes
- − Text rendering is broken and misaligned on the banner
- − Fails to spell the brand name 'Caffè Florian' correctly
- − The primary logo text is illegible
Verdict: Both models failed significantly on typography, with neither generating the correct name 'Caffè Florian'. FLUX.1 [schnell] produced a cleaner, more professional vector graphic, but FLUX.1 [schnell] FP8 was the only one to include the internal details like the steam and the correct establishment year. Ultimately, both are unusable as logos without significant manual correction, but FLUX.1 [schnell] is a slightly better starting point for a design due to its cleaner layout.
Apollo 11: Journey to Tranquility
Text-to-Image“Create a clean, modern vector infographic poster about the Apollo 11 mission. NASA-inspired palette (navy, white, muted red, light gray). Flat-vector style, crisp lines, consistent iconography, subtle gradients only. Steps (stop at landing): 1. Launch (Saturn Vicon) 2. Earth Orbit (Earth + orbit ring icon) 3. Translunar (trajectory arc icon) 4. Lunar Orbit (Moon + orbit ring icon) 5. Descent (lunar module descending icon) 6. Landing (lunar module on the surface icon) Small supporting elements (minimal text): • Crew strip: three silhouette icons with only last names: Armstrong, Aldrin, Collins. • Landing site marker: Moon pin labeled "Tranquility" only. Layout constraints: generous margins, large readable labels, clean background with subtle stars. Vector-only, print-poster look, high resolution.”
FLUX.1 [schnell]
FLUX.1 [schnell] FP8
AI Judge Analysis
FLUX.1 [schnell]
- + Successfully captures the requested color palette with a clean navy background.
- + Effective use of subtle gradients and a minimal vector aesthetic.
- + Good layout that implies a journey from center to surface.
- − Fails to include the specific 6-step chronological sequence requested.
- − Text is largely illegible nonsense and does not follow the specific step labels.
- − Iconography is confusing, with a shuttle-like rocket instead of a Saturn V.
FLUX.1 [schnell] FP8
- + Attempts a clear horizontal sequence that more closely follows the 1 through 6 structure.
- + Legible header text and better adherence to specific terminology like 'Saturn V' and 'Landing'.
- + Includes specific icon elements mentioned in the prompt, such as the trajectory arc.
- − Visual composition is cluttered and segmented poorly.
- − Some icons are overly abstract or do not match the labels provided.
- − Contains significant spelling errors in the secondary text.
Verdict: Both models struggle with the specific technical requirements of a 6-step infographic, but FLUX.1 [schnell] FP8 is the better choice because it actually attempts to list the specific steps and labels requested, whereas FLUX.1 [schnell] creates a more abstract artistic piece. FLUX.1 [schnell] FP8 also features a much bolder and more effective layout for a poster, despite some cluttered elements and spelling errors.
Next steps
Explore each model
FP8 quantized variant of Black Forest Labs' FLUX.1 [schnell] model, offering ~2x faster inference with reduced precision while maintaining high-quality image generation in 4 steps