Distilled version of Black Forest Labs' FLUX.2 [dev] outperforming it at a cheaper price. Developed by fal.ai.
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev] Turbo
27.9
arena score
#3 of 62 in Text-to-Image
Top 3 in Text-to-Image
Skill signature
· Text-to-Image
Stable Diffusion 3.5 Large Turbo
10.7
arena score
#61 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [dev] Turbo
0%
win rate
Ties
0%
Stable Diffusion 3.5 Large Turbo
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Perfect adherence to spatial relationships, placing the book on top and the sphere inside.
- + Highly realistic textures on the wood grain, glass thickness, and book cover.
- + Excellent lighting and depth of field that creates a convincing photographic aesthetic.
- − Small dust artifacts on the glass might be seen as slightly messy, though realistic.
Stable Diffusion 3.5 Large Turbo
- + Clean, vibrant colors and sharp lines.
- + Captures the soft window light effect well with distinct shadows.
- − Failed spatial prompt adherence by placing the red book inside the cube rather than on top.
- − Composition feels more like a 3D render than a real photograph.
- − The plant is awkwardly cropped and positioned compared to the prompt's request for it to be behind.
Verdict: FLUX.2 [dev] Turbo followed the prompt perfectly, correctly placing the book on top of the cube and the sphere inside, while maintaining high photographic realism. Stable Diffusion 3.5 Large Turbo failed to follow the spatial instructions, placing the book inside the cube and creating a more artificial, CGI-like appearance.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent adherence to the 'candid street photo' aesthetic with authentic Japanese urban details.
- + Superior rendering of wet textures, reflections, and rain effects.
- + Very high level of detail in the bicycle mechanics and the man's facial features.
- − None notable, as it perfectly captures both the subject and the requested photographic style.
Stable Diffusion 3.5 Large Turbo
- + Successfully includes the red bicycle and elderly man.
- − Fails to capture the realism requested, appearing more like a 3D render than a photo.
- − Lacks the motion blur, light rain, and complex reflections specified in the prompt.
- − Anatomical issues with the man's posture and the bicycle's integration with his body.
Verdict: FLUX.2 [dev] Turbo produced a stunningly realistic photograph that adheres to every technical instruction, from the 50mm lens look to the motion blur of passing cars. In contrast, Stable Diffusion 3.5 Large Turbo failed to deliver a photographic result, instead producing a simplified, stylistically flat image with anatomical glitches and a lack of environmental atmosphere.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent adherence to the 'beads in hair' prompt with realistic placement.
- + Superior texture rendering on the leather straps, fabric, and ornate engraving.
- + Highly realistic skin details involving pores, subtle scars, and believable dirt.
- − The torch in the background is slightly distracting and lacks a bit of depth integration.
Stable Diffusion 3.5 Large Turbo
- + Strong color contrast with vibrant greens and golds.
- + Smooth, painterly aesthetic that provides a stylized fantasy look.
- − Failed to include the beads in the hair as requested.
- − Skin and hair textures look overly smooth and 'plastic', lacking the requested lifelike texture.
- − The blood/scars look like flat face paint rather than physical wounds.
Verdict: FLUX.2 [dev] Turbo significantly outperformed Stable Diffusion 3.5 Large Turbo by capturing every specific detail of the prompt, especially the fine textures of the leather, engraving, and hair beads. Stable Diffusion 3.5 produced a much smoother, less detailed image that missed several key descriptors and lacked the 'battle-worn' realism requested.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent typography with clean sans-serif fonts
- + Perfect layout that looks like a functional, professional menu
- + Realistic food photography that matches the specified categories
- − Prices are nonsensical (e.g., $250 for a pizza)
- − Some minor Latin-style filler text for descriptions
Stable Diffusion 3.5 Large Turbo
- + Features a vibrant grid of food photos as requested
- + Clean white background with minimalist spacing
- − Layout is impractical with text cards floating at the bottom
- − Food items look overly stylized or artificial
- − Typographical errors including misspelling 'Mains' as 'Mians' and unreadable menu items
Verdict: FLUX.2 [dev] Turbo significantly outperforms Stable Diffusion 3.5 Large Turbo by creating a cohesive and believable menu design that follows professional graphic design principles. While Stable Diffusion 3.5 creates an interesting artistic arrangement, it fails as a functional menu design, including typos and an illogical flow, whereas FLUX.2 produces a high-quality, ready-to-use template.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent typography with perfect spelling and fiery effects.
- + Perfect adherence to the 'exploded' layout requested in the prompt.
- + Highly photorealistic textures on the food components.
- − The lighting on the food is a bit flat compared to the intensity of the background flames.
Stable Diffusion 3.5 Large Turbo
- + Strong sense of drama and vibrant colors.
- + Creative use of smoke and glowing particles.
- − Failed to include any of the requested text.
- − The burger is not exploded or suspended as requested; it is mostly intact on a skewer.
- − Art style is more illustrative/3D render than photorealistic.
Verdict: FLUX.2 [dev] Turbo succeeded in every aspect of the prompt, including complex typography, a specific 'exploded' layout, and photorealistic textures. Stable Diffusion 3.5 Large Turbo failed to include any text and ignored the directive for an exploded view, resulting in a generic (though visually striking) burger image.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Near-perfect text rendering of all requested menu items and prices.
- + Highly realistic chalk texture with smudge marks and natural handwriting variations.
- + Authentic café atmosphere with realistic depth of field and lighting.
- − The trailing sentence ends abruptly rather than completing the full 'Brown Butter' text from the prompt.
Stable Diffusion 3.5 Large Turbo
- + Clean, high-contrast composition with a modern aesthetic.
- + Balanced layout with symmetrical plant placement.
- − Poor text accuracy with numerous misspellings like 'ocopus' and 'clame'.
- − Text looks like a digital font rather than authentic hand-lettered chalk.
- − Fails to follow the specific date and layout requested in the prompt.
Verdict: FLUX.2 [dev] Turbo followed the prompt with exceptional accuracy, producing a highly realistic chalkboard with readable, natural handwriting and correct pricing. In contrast, Stable Diffusion 3.5 Large Turbo failed significantly on text rendering, producing various spelling errors and a digital-feeling font that ignored the specific requests of the prompt.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent fine details on the spacesuit and horse's mane.
- + Cinematic lighting with a realistic use of lens flares and planetary backdrops.
- + Correct interpretation of a 'horse riding astronaut' conceptually, though it missed the paradoxical phrasing.
- − Failed to follow the specific spatial instruction 'horse on top, not vice versa'.
- − The perspective of the horse's legs feels slightly cramped at the bottom.
Stable Diffusion 3.5 Large Turbo
- + Dynamic composition with a strong sense of movement.
- + Clean, minimalist space aesthetic that leans into the surreal request.
- − Failed to follow the specific spatial instruction 'horse on top, not vice versa'.
- − Anatomical issues where the horse's front legs meet the body.
- − The background earth image is warped and lacks realistic detail compared to Model A.
Verdict: Both models failed the negative constraint/spacial logic of placing the horse on top of the astronaut, choosing instead to render a standard astronaut riding a horse. FLUX.2 [dev] Turbo is the clear winner due to its superior texture work, cinematic lighting, and overall image coherence compared to Stable Diffusion 3.5 Large Turbo.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent photorealism in fur textures and lighting.
- + Properly places the passenger in the back seat as requested.
- + Highly accurate adherence to the woman's bored expression and activity (using phone).
- − The passenger appears to be in the front passenger seat rather than the back seat, despite the prompt.
- − The 'TAX' sign on the roof is slightly cut off and oddly rendered.
Stable Diffusion 3.5 Large Turbo
- + Features a sharp, stylized aesthetic with vibrant colors.
- + Covers the basic capybara driver elements well.
- − The passenger is barely visible and does not appear to be using a phone or following the bored expression prompt.
- − The 'TAXI' text on the hat is mirrored/distorted.
- − The capybara's fur looks more like a 3D render than a photorealistic animal.
Verdict: FLUX.2 [dev] Turbo followed the prompt with much greater attention to detail, capturing the specific bored expression of the businesswoman and the intricate texture of the capybara's fur. Stable Diffusion 3.5 Large Turbo failed to include the passenger's activity and had several text rendering issues 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.”
FLUX.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Perfect adherence to all text requirements, including the title, banner, and date/location details.
- + Highly detailed and moody vintage gothic aesthetic with rich textures on the parchment and trees.
- + Excellent composition that feels like a professional, polished invitation design.
- − The glowing eyes on the bats are a bit small and easy to miss.
Stable Diffusion 3.5 Large Turbo
- + Clean graphic design with a clear central focal point.
- + Good use of negative space for the center of the image.
- − Failed to include almost all requested text, including the scroll banner, specific date, time, and location.
- − The style feels more like a modern clip-art collage than a 'vintage gothic cinematic' poster.
- − The background lacks the 'moody night sky' and 'dark parchment' requested, appearing quite bright and flat.
Verdict: FLUX.2 [dev] Turbo followed the prompt with near-perfect accuracy, successfully rendering every line of text, the scroll banner, and the specific event details within a beautifully atmospheric gothic scene. Specifically, Stable Diffusion 3.5 Large Turbo failed to include the majority of the required information, including the date, time, and location, and opted for a much simpler, less cinematic style. FLUX.2 is the clear winner for its superior text rendering and adherence to the complex aesthetic instructions.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent typography with perfect spelling of 'JAPAN' and 'SUSHI'.
- + Sophisticated texture rendering for the rice and fish following PBR materials requirement.
- + Clean, professional layout that perfectly matches the requested diorama style.
- − The perspective is a side oblique rather than a strict 45-degree isometric view.
Stable Diffusion 3.5 Large Turbo
- + Strong isometric composition with a clear 3D miniature feel.
- + Vibrant colors that lean into the 'cartoon scene' aspect of the prompt.
- − Spelling error in 'SUSHI' (rendered as 'SIIHI').
- − The flag icon is incorrect, resembling the flag of Poland rather than Japan.
- − The background is not the requested solid light blue.
Verdict: FLUX.2 [dev] Turbo produced a much higher quality image with perfect text rendering and realistic textures that satisfy the PBR materials requirement. While Stable Diffusion 3.5 Large Turbo captured the isometric perspective well, it failed on basic text accuracy and produced a flag that did not match the subject matter.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Successfully includes all four requested animals (dog, cat, bunny, and fox).
- + Exceptional realism in fur texture and lighting, including the god rays and dew sparkles mentioned in the prompt.
- + Dynamic and coherent composition that captures the 'tumbling' and 'chasing' action well.
- − The kitten's anatomical posing is slightly awkward during the jump.
- − Some butterflies appear to have minor wing clipping issues against the background.
Stable Diffusion 3.5 Large Turbo
- + Vibrant colors and high-contrast lighting create a striking visual.
- + Clean, cute character designs with large expressive eyes.
- − Failed to include all four animals, missing the baby bunny.
- − Lacks photorealism, appearing more like a 3D digital illustration or 'plastic' style.
- − Anatomy of the fox/cat hybrid on the right is ambiguous, and the paws of the puppy are distorted.
Verdict: FLUX.2 [dev] Turbo followed the complex prompt perfectly, including all four specific animals and delivering a masterclass in realistic lighting and texture. Stable Diffusion 3.5 Large Turbo failed on both prompt adherence, missing the rabbit, and visual quality, providing a stylized image rather than the requested photorealistic scene. FLUX.2 also captured the playful movement described, whereas Stable Diffusion produced a static, posed shot.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Perfect adherence to text, including the correctly placed grave accent on 'Caffè'.
- + Excellent vintage texture and balanced, minimalist vector composition.
- + Very clean typography and professional layout of the banner elements.
- − The steam lines are a bit generic compared to the rest of the illustration.
Stable Diffusion 3.5 Large Turbo
- + Dynamic lighting and glossy finish give the illustration more depth.
- + Interesting creative choice to combine the cloche dome with a coffee mug shape.
- − Significant spelling errors in the brand name ('Caffeé Florin').
- − The text rendering on the banner is shaky and lacks the clean vector aesthetic requested.
- − The cloche/mug hybrid is a bit cluttered compared to the minimalist request.
Verdict: FLUX.2 [dev] Turbo significantly outperformed the other model by accurately rendering the specific text 'Caffè Florian' and maintaining a clean, professional vector aesthetic. Stable Diffusion 3.5 Large Turbo struggled with spelling and produced a more cluttered, less balanced emblem that failed the minimalist requirement.
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.2 [dev] Turbo
Stable Diffusion 3.5 Large Turbo
AI Judge Analysis
FLUX.2 [dev] Turbo
- + Excellent text rendering with almost no spelling errors.
- + Specifically followed the 6-step orbital sequence requested in the prompt.
- + Used a clear, modern flat-vector style with appropriate NASA-inspired colors.
- − Includes a redundant 'Saturn Iicon' label taking the prompt instruction too literally.
- − Non-linear layout makes the flow of information slightly confusing to follow.
Stable Diffusion 3.5 Large Turbo
- + Strong aesthetic appeal with a high-quality vintage minimalist design.
- + Consistent vector style with a very professional layout and balance.
- − Failed significantly on text rendering, featuring gibberish and misspellings like 'Apoll.o'.
- − Missing most of the requested steps and specific icons like the Saturn V rocket.
Verdict: FLUX.2 [dev] Turbo is the clear winner because it accurately followed the complex instructional prompt, rendered all requested steps, and produced legible, correct text. Stable Diffusion 3.5 Large Turbo created a visually pleasing artistic poster but failed to include the specific information and icons requested, and suffered from severe text artifacts.
Next steps
Explore each model
Distilled version of SD 3.5 Large that generates high-quality images in just 4 steps, offering faster inference and reduced costs