Fast distilled version of Black Forest Labs' FLUX.2 [dev] optimized for speed and cost efficiency.
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev] Flash
27.0
arena score
#7 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Qwen Image 2512
22.9
arena score
#30 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [dev] Flash
50.0%
win rate
Ties
50.0%
Qwen Image 2512
0.0%
win rate
50.0%
50.0% ties
0.0%
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent photorealistic textures on the wooden table and red book.
- + Highly accurate glass physics, including subtle dust and realistic reflections.
- + Perfect adherence to all spatial requirements of the prompt.
- − The sphere appears more like a translucent marble than a solid sphere, though it remains blue.
Qwen Image 2512
- + Captures all requested elements correctly.
- + Pleasant soft lighting from the left.
- − The glass physics are slightly inconsistent, with the back edge of the cube appearing warped.
- − The green plant is visible through the glass but less naturally integrated than in the other image.
Verdict: Both models followed the prompt perfectly, but FLUX.2 [dev] Flash produced a significantly more photorealistic result with superior textures and lighting. Qwen Image 2512 is a strong contender but has slight distortions in the glass geometry and less detail in the tabletop wood grain.
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] Flash
Qwen Image 2512
0% wins
100% ties
0% wins
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent adherence to the 'repairing' action with tools on the ground.
- + Effective motion blur on background vehicles and realistic rain textures.
- + High level of detail in skin texture and aged hands.
Qwen Image 2512
- + Strong cinematic lighting and high-quality bokeh.
- + Good facial anatomy and realistic skin texture.
- + Captures the light rain atmosphere well.
- − The subject is posing/staring at the camera rather than 'repairing' the bicycle.
- − The bicycle's structure has some logical errors around the chain guard and frame.
- − Misses the specific 'motion blur' requested for the background cars.
Verdict: FLUX.2 [dev] Flash followed the prompt much more accurately, showing the man actively engaged in the task of repairing the bike with tools visible on the ground and requested motion blur on the cars. Qwen Image 2512 produced a high-quality portrait, but the subject is simply sitting with the bike and looking at the camera, which misses the 'repairing' and 'candid' aspects of the prompt.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent execution of engraved plate armor detail.
- + Strong symmetry and clear bokeh sparks.
- + High-fidelity rendering of leather strap textures and buckles.
- − The 'scars' look more like fresh surface blood than battle-worn scarring.
- − The face appears a bit youthful and clean for a 'battle-worn' character.
Qwen Image 2512
- + Superior facial depth with believable weathered skin and scarred texture.
- + Much better lighting integration with torchlight creating a warm, directional glow.
- + The hair braiding and bead integration feels more organic.
- − The bokeh sparks are slightly less defined than in the other model.
- − The transition between the neck and the armor is a bit cluttered.
Verdict: Qwen Image 2512 captures the 'battle-worn' aesthetic much better, with realistic skin texture and cinematic lighting that properly reflects the torchlight. While FLUX.2 [dev] Flash has incredibly sharp armor details and technical clarity, it feels more like a clean studio portrait compared to the gritty, atmospheric realism of Qwen Image 2512.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent photo realism and lighting on the food items.
- + Strong alignment with the request for specific categories like appetizers and mains.
- + Includes recognizable pricing and structured layout.
- − Text rendering is very poor with significant gibberish and artifacts.
- − The food grid is repetitive, featuring almost exclusively pizzas.
Qwen Image 2512
- + Successfully creates a more diverse food grid as requested.
- + The layout is very clean with a professional, cohesive typographic style.
- + The vibrant accents are used more intentionally to define categories.
- − The text is entirely illegible mock-language.
- − The categorization logic is confused, listing 'Piesmanets' instead of clear pizzas/mains.
Verdict: Both models struggled with legible text, but Qwen Image 2512 followed the prompt regarding a diverse food grid much better than FLUX.2 [dev] Flash, which mostly showed different angles of the same pizza. Qwen Image 2512 also exhibited a more balanced and professional layout that feels like a real modern menu design, whereas FLUX.2 had more technical artifacts around the top text.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent text rendering with clean fiery outlines
- + Perfect adherence to all requested text elements
- + Clean, high-quality photorealistic rendering of food textures
- − The 'exploded' effect is a bit static compared to the kinetic debris in the other image
- − The background embers are somewhat subtle
Qwen Image 2512
- + Highly dynamic 'exploded' effect with many flying debris particles
- + Bold, impactful typography with a strong 3D fiery effect
- + Vibrant colors and a strong sense of lighting
- − Missed the word 'TIME' in the 'LIMITED TIME ONLY' phrase
- − The cheese and patty textures look slightly more plastic-like than real meat
Verdict: Both models followed the visual brief successfully, creating high-impact ad layouts. FLUX.2 [dev] Flash is the winner because it followed the text prompt precisely, whereas Qwen Image 2512 omitted a word from the secondary message. FLUX.2 also provided a more realistic texture on the hamburger patty.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent chalk texture throughout all text
- + Correct spelling for all menu items
- + Natural smudges and board texture add to the realism
- − The handwriting style is a bit inconsistent between the bold title and thinner body text
Qwen Image 2512
- + Very consistent and aesthetically pleasing cursive handwriting style
- + Perfect layout and centering of the text on the board
- + Extremely crisp text rendering
- − Spelling error in 'Risitto' (should be Risotto)
- − The chalk texture feels slightly too clean and digital compared to Model A
Verdict: Both models followed the complex text-heavy prompt impressively well. FLUX.2 [dev] Flash captures the most realistic chalk artifacts and achieves perfect spelling, while Qwen Image 2512 offers a more beautiful calligraphic style and better composition but fails on the spelling of 'Risotto'.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [dev] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + High cinematic visual quality with atmospheric background detail.
- + Excellent lighting and color palette that feels surreal yet grounded.
- + Consistent rendering of the astronaut's gear and horse's mane.
- − Failed the spatial prompt logic by placing the astronaut on top of the horse.
Qwen Image 2512
- + Good clarity on the astronaut's face and equipment lighting.
- + Sharp texture on the horse's coat and leather saddle.
- − Failed the spatial prompt logic by placing the astronaut on top of the horse.
- − The composition is less cinematic compared to the other model.
Verdict: Both FLUX.2 [dev] Flash and Qwen Image 2512 completely failed the specific negating instruction to have the 'horse on top' of the astronaut, instead providing the cliché astronaut-riding-horse image. FLUX.2 [dev] Flash is the slightly better image overall due to its superior cinematic composition, detailed background, and better integration of the subjects into the environment.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent adherence to the 'bored' expression for the passenger
- + Captures the New York taxi aesthetic well with the 'TAX' sign on top
- + High level of photographic realism and lighting
- − The passenger is sitting in the front passenger seat instead of the 'back seat' as requested
Qwen Image 2512
- + Correctly places the passenger in the back seat as specified in the prompt
- + High quality rendering of the capybara's fur and the taxi driver hat
- + Realistic depth of field and street lighting
- − The passenger's expression looks more 'annoyed' or 'angry' than specifically 'bored'
- − The capybara's paws look slightly more like human hands in gloves than actual paws
Verdict: While both models produced high-quality, professional results, Qwen Image 2512 followed the spatial instructions more accurately by placing the passenger in the back seat. FLUX.2 [dev] Flash captured the 'bored' expression better but failed the positional prompt by putting the businesswoman in the front seat next to the driver.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Perfect spelling on all requested text elements.
- + Excellent border integration with the parchment roll aesthetic.
- + Rich, moody lighting that enhances the gothic theme.
- − Includes a small line of garbled text ('Hurk: 0369c') that was not in the prompt.
Qwen Image 2512
- + Dynamic composition with swirling vines on the pumpkin.
- + Clear, legible event details at the bottom.
- + Atmospheric lighting and color palette.
- − Includes a spelling error in the main title ('Hallowern').
- − The scroll banner is slightly warped at the left edge.
Verdict: FLUX.2 [dev] Flash is the winner as it correctly spells the main title, unlike Qwen Image 2512 which output 'Hallowern'. While FLUX.2 [dev] Flash added one line of hallucinated text, its overall execution of the gothic typography and elegant parchment border makes it a much more functional invitation.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Perfect text rendering and layout placement.
- + Strong PBR material qualities, especially on the wooden base and fish texture.
- + Very clean and minimalist aesthetic that matches the 'ultra-clean' requirement.
- − The sushi construction is slightly physically illogical (nigiri-style toppings on a roll-style base).
Qwen Image 2512
- + Excellent 'cartoon' diorama style with more variety in sushi types.
- + Creative typography that fits a 3D cartoon theme better than plain black text.
- + Better adherence to the 'miniature 3D cartoon' style requested.
- − The 'JAPAN' text has slight alignment issues with the 'SUSHI' text below it.
- − The lighting is a bit flatter compared to the realistic shadows in Model A.
Verdict: FLUX.2 [dev] Flash produces a cleaner, more realistic image with superior text rendering and high-quality materials. However, Qwen Image 2512 captures the 'cartoon diorama' spirit much better, offering a more interesting scene with varied sushi and a stylized base, even if the text is slightly less polished.
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] Flash
Qwen Image 2512
100% wins
0% ties
0% wins
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent depiction of movement with the puppy and kitten pouncing.
- + Stronger adherence to the 'god rays' and 'dew sparkles' requirement with visible water droplets on the grass.
- + Very high fur detail and natural lighting integration.
- − Includes an extra fifth animal (a second bunny) not requested in the prompt.
Qwen Image 2512
- + Perfect adherence to the count and species of animals requested.
- + Centrally balanced and aesthetically pleasing composition.
- + Excellent rendering of the golden retriever's face and fur.
- − Static posing that doesn't fully capture 'playfully chasing' or 'tumbling' as well as Model A.
- − The fox's ears and proportions look slightly off compared to a real kit.
Verdict: FLUX.2 [dev] Flash captures the dynamic action of the prompt much better, showing animals pouncing and playing, but it fails the logic test by adding a fifth animal. Qwen Image 2512 provides a cleaner, more accurate set of animals and a highly professional composition, though it feels more like a static portrait than an action scene. Qwen Image 2512 is the preferred overall winner for following the specific animal list while maintaining high visual quality.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent typography with proper accent mark usage.
- + Follows the minimalist vector emblem style perfectly.
- + Great color balance and clean layout.
- − The steam is very simple and could be more integrated.
- − Banner placement is slightly overlapping the cloche base in a heavy way.
Qwen Image 2512
- + Dynamic and beautifully rendered steam swirls.
- + High level of detail in the cloche highlights and textures.
- + Strong retro aesthetic with nice shading.
- − The 'Est. 1720' text is slightly cramped within the banner curve.
- − Less minimalist than requested, leaning more towards an illustration than a vector logo.
Verdict: Both models followed the prompt well, but FLUX.2 [dev] Flash produced a more authentic logo that adheres to the 'minimalist' and 'vector emblem' keywords. Qwen Image 2512 created a beautiful vintage illustration, but the complexity of the steam and shading makes it less practical as a logo compared to the clean design of FLUX.2 [dev] Flash.
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] Flash
Qwen Image 2512
AI Judge Analysis
FLUX.2 [dev] Flash
- + Excellent typography with very accurate spelling for most labels.
- + Strong adherence to the NASA-inspired color palette and vector style.
- + Includes creative extra details like the crew portraits and landing site marker.
- − Step numbering is absent, leading to a cluttered and non-linear flow.
- − Text artifacts under the launch vehicle are messy.
Qwen Image 2512
- + Features a clear numbered sequence for the mission steps.
- + Cleaner vertical composition that feels more like a traditional poster.
- + Good adherence to the requested flat vector style.
- − Several spelling errors in key labels like 'Desceeint' and 'Translaurtcoit'.
- − Includes prompt text 'Steps stop at landing' as literal text on the poster.
- − Step 1 (Launch) is not explicitly labeled with text.
Verdict: FLUX.2 [dev] Flash produces a much more visually polished and professional infographic with superior text rendering and better illustration quality. While Qwen Image 2512 does a better job of numbering the steps sequentially, its frequent spelling mistakes and the inclusion of prompt text as literal labels make it less effective overall.
Next steps
Explore each model
Improved version of Alibaba's Qwen image model with better text rendering, finer natural textures, and more realistic human generation.