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
19.0
arena score
#47 of 62 in Text-to-Image
Skill signature
· Text-to-Image
OmniGen v2
16.9
arena score
#57 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.1 [schnell] FP8
0%
win rate
Ties
0%
OmniGen v2
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent handle on complex light refraction within the glass.
- + High level of realistic detail in the wood grain and sphere texture.
- + Accurate interpretation of 'partially visible through the glass' for the plant.
- − The glass object is more of a tall rectangular prism than a strict cube.
- − Includes an unnecessary internal shelf that makes the physics look a bit odd.
OmniGen v2
- + Follows the 'cube' geometry much more accurately than Model A.
- + Clean, minimalist composition with vibrant colors.
- + Accurately places the ball inside and the book on top.
- − The plant is behind the cube but doesn't show any refraction or visibility through the glass, missing a specific prompt detail.
- − The physics of the book sitting on the thin edges of an open-top glass container looks slightly unstable.
Verdict: OmniGen v2 followed the geometric instructions better by providing a true cube, whereas FLUX.1 [schnell] FP8 produced a rectangular prism. however, FLUX.1 [schnell] FP8 captured the complex rendering of the plant seen through the glass and superior lighting effects, making it the more visually impressive image despite the shape minor inaccuracy.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent depiction of wet pavement and bokeh light reflections
- + Realistic skin texture and age-appropriate physical features
- + Captures an 'imperfect' cinematic framing that feels candid
- − Lack of visible rain streaks despite wet ground
- − The man appears more Southeast Asian/Latino than Japanese
OmniGen v2
- + Successfully renders visible rain falling in the air
- + Strong character design that fits the 'Japanese elderly man' prompt perfectly
- + Clean composition with clear reflections
- − The bike anatomy is slightly warped near the pedals and chain guard
- − Missing the requested 'motion blur' from passing cars
- − Skin texture looks somewhat smoothed and less realistic
Verdict: FLUX.1 [schnell] FP8 wins on overall realism, lighting, and cinematic feel, perfectly capturing the wet city atmosphere and shallow depth of field. However, OmniGen v2 adhered better to the specific character ethnicity and the visible rain requirement, though its lack of motion blur and slightly artificial skin texture make it feel less like a candid photo.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Exceptional high-frequency detail in the skin texture and eyes
- + Strong cinematic lighting with realistic warm highlights and cool shadows
- + Intricate engraving on the armor and leather straps
- − Braided hair requirement is represented very minimally as thin strands rather than distinct braids
- − The dirt and scars on the skin are very subtle compared to the prompt's focus
OmniGen v2
- + Excellent adherence to the 'hair braided with small beads' requirement
- + Clearly visible dirt and scars on the facial skin
- + Distinct bokeh sparks and warm firelight effect
- − Slightly soft/filtered appearance on the skin compared to the high-detail requirement
- − Plate armor feels somewhat flatter and less 'worn' than Model A
Verdict: FLUX.1 [schnell] FP8 delivers a more hyper-realistic and intense 'battle-worn' feeling with superior texture on the metal and skin, though it missed the specific braiding style requested. OmniGen v2 adhered much better to the specific hair and bead details and included clearer dirt/scars, but the overall image quality looks slightly more digital/artificial. FLUX.1 [schnell] FP8 is the narrow winner for its professional-grade texture and lighting that better captures the mood of a paladin.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent structure that closely resembles a real two-page menu layout
- + Very clean and readable typography and section headers
- + Consistent photography style across all food items
- − Several spelling errors in headings like APPTIZERS and SECCER
- − A large portion of the text is illegible placeholder gibberish
OmniGen v2
- + Stronger use of vibrant accents and colorful block backgrounds
- + Higher resolution and clarity in the food photography
- + Includes larger feature images which add visual hierarchy
- − Layout is somewhat disorganized with repetitive section headers like PIZZA/PIZZZAN
- − Major spelling errors in main titles and headers
- − Sections are less clearly defined than Model A
Verdict: FLUX.1 [schnell] FP8 provides a much more convincing menu layout with professional spacing and a logical grid, although it suffers from several spelling mistakes in the headers. OmniGen v2 has better individual image quality and punchier colors, but the repetitive headings and disorganized document structure make it less effective as a design piece. FLUX.1 [schnell] FP8 is the preferred choice for its superior composition and professional feel.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Successfully rendered the 'exploded' burger effect with mid-air components.
- + Includes realistic elements like steam/smoke and scattered ingredients for texture.
- + Followed the fiery background and glowing embers requirement well.
- − Serious spelling errors including 'LIIMITED' and 'NEEY'.
- − Incorrect price representation with the large '€69' text.
- − The burger is largely assembled rather than a full exploded view of all layers.
OmniGen v2
- + High-quality graphic design aesthetic suitable for an advertisement.
- + Accurate text rendering for 'MAGIC BURGER' and the price '6.99'.
- + Clean, vibrant fiery effect with good composition.
- − Failed the 'exploded burger' prompt, showing a fully assembled static burger.
- − The 'LIMITED TIME ONLY' text is cut off on the left side.
- − Missing the currency symbol '€' next to the price.
Verdict: FLUX.1 [schnell] FP8 better captured the 'exploded' concept and cinematic lighting, but suffered from significant text hallucinations and pricing errors. OmniGen v2 produced a cleaner graphic and more accurate titles, but failed the primary structural requirement of the prompt by presenting a static, non-exploded burger. FLUX.1 is slightly better for adhering to the core visual concept, despite the poor spelling.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + The font style looks very much like authentic chalk handwriting on a board.
- + The wood frame and lighting create a realistic 'cozy café' atmosphere.
- + The date text is accurately rendered as requested.
- − Has significant spelling errors and repetitive lines (e.g., 'Grillled', 'cokiies').
- − Fails to follow the request for elegant cursive in the title.
OmniGen v2
- + Attempts to include more of the prompt's requested text items.
- + Captures the smudge marks typically found on real chalkboards.
- + The layout is well-centered with clear price placement.
- − Severe spelling errors such as 'SPECALS', 'Octapus', and 'Lemont'.
- − Text becomes cluttered and visually messy in the middle section.
- − The title font feels somewhat stylized/digital rather than purely handwritten cursive.
Verdict: Both models struggled significantly with spelling accuracy and the specific cursive instruction. FLUX.1 [schnell] FP8 produced a much more realistic photograph with convincing chalk textures and café lighting, whereas OmniGen v2 felt more like a flat graphic. FLUX.1 [schnell] FP8 is preferred for its superior aesthetic quality and more natural handwriting style, despite the repetition in the 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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Successfully followed the difficult logic of placing the horse on top of an object/being.
- + Superior cinematic composition with a grander sense of scale and detail.
- + Captures the surreal atmosphere requested in the prompt.
- − The astronaut is represented as a mechanical box/suit rather than a recognizable human figure.
- − Visual artifacts present where the two horse bodies intersect.
OmniGen v2
- + Clear and understandable rendering of an astronaut and a horse.
- + High contrast and clean lines in the subject.
- − Completely failed the negative constraint/logic of having the horse on top.
- − Generic illustrative style rather than cinematic and surreal.
- − Lacks the epic scale and lighting of the first image.
Verdict: FLUX.1 [schnell] FP8 is the clear winner because it successfully attempted the surreal logic of the prompt ('horse on top'), even if it interpreted the astronaut as a piece of equipment. OmniGen v2 ignored the specific instruction to reverse the traditional rider/mount roles, providing a standard astronaut-on-horse image that lacks the requested cinematic quality.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent adherence to the 'taxi' theme with clear 'TAXI' text on the cap.
- + Accurate fur textures and realistic capybara anatomy for the driver.
- + Correct seating arrangement with the driver in the front and passenger in the back.
- − The passenger is strangely holding two phones.
- − The capybara's paw placement on the steering wheel is slightly awkward compared to human-like steering.
OmniGen v2
- + High resolution and very clean, sharp image quality.
- + Professional looking business attire for both the driver and passenger.
- − Failed the spatial layout by placing human hands on the steering wheel instead of the capybara's paws.
- − The passenger appears to be in the front passenger seat rather than the back seat as requested.
Verdict: FLUX.1 [schnell] FP8 correctly followed the complex instruction for the driver and passenger seating arrangement, although the passenger has an extra phone. OmniGen v2 produced a higher-fidelity image but failed significantly on the logic of the scene, giving the capybara driver human hands and placing the passenger in the front seat.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Features a more cinematic and three-dimensional jack-o-lantern.
- + Captures the moody, dark parchment aesthetic well.
- + The twisted trees create a strong gothic atmosphere.
- − Significant text failures with multiple typos and gibberish.
- − The layout of the event details is messy and confusing.
- − The lower scroll banner contains broken English.
OmniGen v2
- + Much cleaner title rendering and overall font legibility.
- + Accurately includes elements like spider webs and thorns in the border.
- + Better adherence to specific dates and locations in the text.
- − The jack-o-lantern is a silhouette rather than a glowing 3D object.
- − The scroll banner text is partially garbled and strangely placed.
- − The visual style is a bit more 'cartoonish' than 'cinematic'.
Verdict: OmniGen v2 is the superior choice because it maintains much higher text accuracy, correctly rendering the date and location, whereas FLUX.1 [schnell] FP8 produced significant typos and gibberish. While FLUX.1 [schnell] FP8 captured the lighting and 'cinematic' mood better, it failed the fundamental requirement of an invitation to communicate specific event details.
Isometric Miniature Diorama Scenes
Text-to-Image“Create a clear, 45° top-down isometric miniature 3D cartoon scene of Japan's signature dish: sushi, with soft refined textures, realistic PBR materials, gentle lighting, on a small raised diorama base with minimal garnish and plate. Solid light blue background. At top-center: 'JAPAN' in large bold text, 'SUSHI' below it, small flag icon. Perfectly centered, ultra-clean, high-clarity, square format.”
FLUX.1 [schnell] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent 3D miniature styling with soft, refined textures.
- + Great lighting and shadow work on the base and food.
- + Includes a clever integration of the Japanese flag within the text.
- − Failed the text prompt, repeating 'JAPAN' incorrectly and omitting 'SUSHI'.
- − Background is very pale, almost white rather than light blue.
OmniGen v2
- + Perfect adherence to text instructions, including both words and the flag icon.
- + Stronger 45° isometric diorama composition as requested.
- + Vibrant colors and clean graphic design quality.
- − The flag used is not the flag of Japan.
- − Perspective of the sushi pieces is slightly inconsistent with the tray.
- − Texture is more 2D-illustrative than 'realistic PBR'.
Verdict: OmniGen v2 is the winner because it successfully followed the complex text instructions and the isometric diorama request, whereas FLUX.1 [schnell] FP8 failed to include the word 'SUSHI' and garbled the text. While FLUX.1 produced superior lighting and textures, OmniGen v2 delivered a more accurate representation of the full prompt's layout and content requirements.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent fur texture and realistic lighting with clear god rays.
- + Captures a high level of detail in the eyes and floral foreground.
- + Includes a wider variety of animals that better match the prompt species.
- − Failed to include a distinct bunny, opting for multiple fox/kitten hybrids.
- − The composition is a bit crowded with overlapping subjects.
OmniGen v2
- + Strong sunrise lighting and clear, bright colors.
- + Features a very clear golden retriever puppy and kitten.
- − The style is heavily stylized/illustrative rather than 'hyper-photorealistic'.
- − Missing the bunny and red fox kit, showing fewer animals than requested.
- − Butterflies and animals look like flat clip-art layers rather than being integrated.
Verdict: FLUX.1 [schnell] FP8 is the clear winner as it provides a much more realistic and detailed scene that aligns with the 'hyper-photorealistic' part of the prompt. OmniGen v2 produced a generic, cartoonish illustration that failed to include all requested animals and lacked the sophisticated fur textures and atmospheric depth seen in the FLUX image.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Excellent color palette and retro texture
- + Intricate Victorian-style architectural interpretation of a cloche
- + Successfully includes all prompt elements like the 'Est. 1720' banner
- − Major spelling errors in the main text ('AFe FLAMILAN')
- − Typography is cluttered and overlapping
OmniGen v2
- + Near-perfect spelling of the brand name
- + Clean, minimalist vector aesthetic that fits the 'logo' prompt
- + Clear representation of the cloche dome and steam
- − Slightly missed the space in 'Caffè Florian'
- − The composition of the banner and the bottom curve is a bit disconnected
Verdict: While FLUX.1 [schnell] FP8 captures the 'vintage' and 'warm texture' vibes much better with more artistic flair, it completely fails the text rendering, outputting 'AFe FLAMILAN'. OmniGen v2 provides a much more usable logo with clean lines and readable text, although it leans slightly more modern than requested.
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] FP8
OmniGen v2
AI Judge Analysis
FLUX.1 [schnell] FP8
- + Follows the specific request for 6 steps more closely
- + Captures the NASA-inspired flat vector aesthetic accurately
- + Text rendering of the main title 'APOLLO 11' is correct
- − Icons do not always match the specific text labels (e.g., Saturn V icon is just a circle)
- − The layout becomes a bit cluttered and nonsensical toward the bottom
OmniGen v2
- + Very clean and balanced composition
- + Icons are more visually distinct and professional in a vector style
- + Excellent use of the requested color palette
- − Failed the primary subject by writing 'APOLO 17' instead of 'APOLLO 11'
- − Only includes three main info blocks, ignoring the 6-step requirement
- − Spelling errors in the secondary text are significant
Verdict: FLUX.1 [schnell] FP8 followed the sequence of instructions much better, providing a 6-step layout as requested and getting the mission name right. OmniGen v2 produced a more visually polished vector design but failed significantly on factual prompt adherence, including the wrong mission number and ignoring the step-by-step requirements.
Next steps
Explore each model
Unified multimodal model for text-to-image generation, instruction-guided image editing, personalized generation, and virtual try-on