Black Forest Labs' distilled 9 billion parameter image generation model with sub-second inference and multi-reference support
Settled by community votes across 17 shared challenges, with an AI judge weighing in on each.
FLUX.2 [klein] 9B
#13 of 32 in Image Editing
Z-Image Turbo
#12 of 62 in Text-to-Image
Where the votes landed
FLUX.2 [klein] 9B
50.0%
win rate
Ties
0.0%
Z-Image Turbo
50.0%
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.2 [klein] 9B
- + Excellent photorealism and texture on the wooden table.
- + Highly realistic glass refraction and light play.
- + Beautiful soft lighting that matches the requested window light.
- − The glass object is more of a hollow jar shape than a sharp geometric cube.
- − The blue sphere appears to be floating mid-air inside the glass rather than resting on the bottom.
Z-Image Turbo
- + Perfect geometric cube shape with thin, realistic glass panes.
- + The blue sphere is realistically grounded on the floor of the cube with a clear reflection.
- + Great composition that accurately places the plant behind the glass.
- − The textures are slightly smoother and less detailed than FLUX.2.
- − The lighting is a bit flatter compared to the more cinematic bloom in Image A.
Verdict: Both models followed the prompt instructions perfectly, including the specific spatial arrangement of the plant and objects. While FLUX.2 [klein] 9B offers superior photographic quality and lighting, Z-Image Turbo captures the 'cube' geometry and the physics of the sphere resting on a surface much more accurately.
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.2 [klein] 9B
- + Excellent skin texture and hyper-realistic facial details
- + Strong atmosphere with visible rain and beautiful wet pavement reflections
- + Follows 'repairing' prompt accurately with tools and squatting posture
- − The white car in the background is static rather than having motion blur
- − Minor anatomy inconsistencies in how the hands grip the frame
Z-Image Turbo
- + Successfully captures motion blur on the passing vehicle
- + Good composition with the subject centered against a busy street
- − Subject is just standing with the bike rather than 'repairing' it
- − Human anatomy is warped, particularly the hands and the scale of the forearms
- − Lacks the skin texture and depth requested in the prompt
Verdict: FLUX.2 [klein] 9B is the superior image by a wide margin, delivering a cinematic and highly detailed portrait that accurately depicts the 'repairing' action. While Z-Image Turbo captured the motion blur of the car better, it failed on the specific action and suffered from significant anatomical distortions in the subject's arms and hands.
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.2 [klein] 9B
- + Outstanding detail on the ornate engraving and leather texture
- + Very sharp and lifelike skin textures including realistic scars
- + Excellent lighting consistency with the warm torchlight affecting all surfaces
- − The braids are a bit stiff and symmetrical, appearing slightly unnatural
Z-Image Turbo
- + Good atmospheric mood with localized lighting from the torch
- + Interesting character design and pose
- + Successful implementation of the beaded hair and bokeh sparks
- − Noticeable anatomy issues with the eyes looking in different directions
- − The armor engraving is significantly less detailed than Model A
- − Texture on leather and cloth is somewhat soft and lacks definition
Verdict: FLUX.2 [klein] 9B delivers a superior image with significantly higher fidelity in textures and more accurate facial features. Z-Image Turbo has a nice atmospheric quality but suffers from 'lazy eye' and lacks the sharp, intricate detail requested for the armor and leather straps.
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.2 [klein] 9B
- + Successfully included all requested sections (Appetizers, Pizza, Mains).
- + Features clear, bold sans-serif fonts for the headers.
- + Excellent photo quality for the food dishes.
- − Text is highly garbled and overlaps significantly in the header ('Resteuranrbeteina').
- − Header assignments are redundant (duplicates 'Appetizers' and 'Mains').
- − The lower section layout is cluttered and difficult to read.
Z-Image Turbo
- + Stronger visual grid layout that feels more modern and structured.
- + Better use of vibrant orange accents for a cohesive brand look.
- + Higher variety in food items shown in the grid.
- − Confused 'Mains' with 'Mans' and 'Section' with 'Setiiion'.
- − Logic of sections is confusing (Appetizers section contains pizza photos).
Verdict: Z-Image Turbo is the superior choice because it captures the essence of a modern minimalist menu with a clean, professional grid and vibrant accents. While FLUX.2 [klein] 9B followed the prompt's section requirements more accurately, the execution is hampered by significant text overlapping and poor use of space, whereas Z-Image Turbo's design is aesthetically more professional and balanced despite minor spelling errors.
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.2 [klein] 9B
- + Excellent text integration with complex fiery textures across all prompted phrases.
- + Highly detailed food textures, especially on the meat patties and fresh lettuce.
- + Dynamic sense of motion with flying sauce and embers that follow a central explosion theme.
- − The 'exploded' effect is slightly conservative, with components still mostly touching.
Z-Image Turbo
- + Strong vertical composition and bright, legible text.
- + Effective use of glowing embers and steam/fire in the background.
- − The burger is not 'exploded' or suspended in mid-air pieces; it is a solid stack floating.
- − The meat patties have an unnatural, slightly repetitive pebbled texture.
- − The lighting on the starburst feels more like a flat graphic than an integrated 3D element.
Verdict: FLUX.2 [klein] 9B is the clear winner as it more accurately follows the 'exploded' and 'dispersed' instruction and handles the complex fiery text textures with much greater detail and photorealism. Z-Image Turbo produces a standard floating burger advertisement, missing the dynamic deconstructed feeling requested by the prompt.
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.2 [klein] 9B
- + Excellent chalk texture with realistic smudges and ghosting on the board
- + Successfully completed the truncated prompt for 'Brown Butter Chocolate Chip Cookies'
- + Displays a more elegant cursive-leaning style that matches the prompt requirement
- − Includes a spelling error 'Mushrropom'
- − The wooden frame has a slight perspective misalignment with the background café
Z-Image Turbo
- + Features very clear and legible handwriting with good spacing
- + Accurately rendered all requested menu items and prices
- + Includes a clean, centered layout with bullet points for readability
- − Contains a spelling error 'Mustroom'
- − The handwriting style is fairly uniform and lacks the requested 'natural variations in letter size and slight slant'
- − The chalk texture is too clean and looks slightly like a digital chalk font
Verdict: FLUX.2 [klein] 9B is the winner because it captures the authentic 'chalkboard' aesthetic much better, including realistic smudging and a more sophisticated cursive style. While both models had minor spelling errors, FLUX.2 successfully inferred the full name of the cookies despite the prompt being truncated, whereas Z-Image Turbo's output felt a bit more generic and digital.
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.2 [klein] 9B
- + Excellent cinematic rendering of space with diverse celestial bodies
- + High level of detail in the horse's musculature and the astronaut's suit
- + Captures the surreal quality requested through the vibrant, busy composition
- − Generated nonsensical text at the bottom of the image
- − Failed the logic check: the human is riding the horse, not the horse on top of the astronaut
Z-Image Turbo
- + Natural and clean lighting on the horse and astronaut
- + Anatomy of the horse and suit are well-proportioned and realistic
- − Failed the logic check: the horse is being ridden by the human
- − The background is very sparse and lacks the 'highly detailed' and 'cinematic' space elements requested
- − The horse's hind legs are somewhat blurred and poorly defined against the background
Verdict: Both models completely failed the negative constraint and specific logic of the prompt ('horse on top, not vice versa'), with both depicting a standard astronaut riding a horse. However, FLUX.2 [klein] 9B is the superior image due to its much more professional, cinematic composition and highly detailed space environment, whereas Z-Image Turbo produced a very plain and empty background. FLUX.2's inclusion of gibberish text is a minor flaw compared to its overall visual 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.”
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent adherence to the 'back seat' perspective to show the passenger clearly
- + High level of detail in the fur texture and taxi interior
- + Correctly captures the bored, mundane expression of the passenger
- − The woman appears to be sitting in the front passenger seat rather than the back seat
- − Garbled text on the taxi cap
Z-Image Turbo
- + Natural lighting and clean, photorealistic textures
- + Good use of depth of field for the background Manhattan lights
- + Accurate depiction of the capybara's front paws on the steering wheel
- − The passenger is sitting in the front passenger seat, not the back seat
- − The capybara's head looks somewhat pasted onto the body
Verdict: Both models failed to place the passenger in the actual back seat, instead placing them in the front passenger seat next to the driver. FLUX.2 [klein] 9B is the winner due to better composition and a much clearer depiction of the passenger's bored expression and activity, whereas Z-Image Turbo's passenger is somewhat obscured and out of focus.
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.2 [klein] 9B
- + Excellent typography with perfect spelling in all sections.
- + Strong cinematic lighting and integration of atmospheric elements like the moon and mist.
- + Very clean border execution featuring the requested webs and thorns in a cohesive design.
- − The parchment texture is less emphasized than the other model.
Z-Image Turbo
- + Great 'torn parchment' aesthetic which fits the vintage theme well.
- + Clear separation between design elements like the jack-o-lantern and the text.
- − Contains a typo in the location ('Archves' instead of 'Arches').
- − The scroll banner for the 'frights' text is poorly placed at the very top and looks like plain text rather than part of the scroll.
- − The composition feels a bit cluttered with overlapping scrolls and floating bats.
Verdict: FLUX.2 [klein] 9B is the clear winner as it followed all textual instructions perfectly, including the correct spelling of 'The Arches'. Its composition is more polished and professional, with cinematic lighting that naturally integrates the jack-o-lantern and the gothic border. Z-Image Turbo captures the parchment feel well but fails on text accuracy and cohesive layout.
Bald man challenge
Image Editing“Give the person a full, thick head of natural hair with realistic texture, density, and a natural hairline. Preserve facial features and lighting.”
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent adherence to the 'full, thick head of hair' instruction with realistic texture.
- + Perfect preservation of the facial features, glasses, clothing, and background.
- + The lighting on the new hair matches the source environment perfectly.
- − The hairline on the side near the glasses temple is slightly cluttered but still realistic.
Z-Image Turbo
- + Natural-looking buzz cut texture.
- − Failed the core instruction to provide a 'full, thick head' of hair, opting for a buzz cut instead.
- − Removes the subject's glasses which was not requested.
- − Changes the background and the man's facial features slightly, losing the identity of the source subject.
Verdict: FLUX.2 [klein] 9B followed the edit instructions perfectly, adding a convincing head of hair while keeping the original person's identity and environment completely intact. Z-Image Turbo failed most aspects of the prompt by removing the glasses, changing the background, and only adding a short buzz cut rather than a full head of hair.
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.2 [klein] 9B
- + Excellent text rendering with clean, sharp typography
- + Realistic wood texture on the diorama base
- + High level of detail on the sushi rolls and salmon texture
- − The flag icon is incorrect, appearing closer to Yemen than Japan
- − A bit of graininess in the soft shadows on the blue background
Z-Image Turbo
- + Successfully captures the soft, 'cartoon' 3D aesthetic requested
- + Very clean, rounded geometry on the diorama base
- + Good use of vibrant colors and soft lighting
- − Includes the Chinese flag instead of the Japanese flag
- − Text is slightly less crisp and has minor kerning/weight inconsistencies
- − The sushi anatomy is less realistic with the green filling appearing strangely placed
Verdict: Both models struggled with the specific 'Japan flag' requirement, with Flux.2 [klein] 9B providing an abstract tricolor and Z-Image Turbo providing the Chinese flag. Flux.2 [klein] 9B is the winner due to its superior text rendering, more professional 3D layout, and more realistic representation of sushi materials and wood textures while maintaining the isometric style.
Over-the-top cartoon caricature
Editing“Create a caricature of me and my job. Make it exaggerated and humorous, incorporating my profession as a tv show anchor and my love for dogs and hockey.”
AI Judge Analysis
FLUX.2 [klein] 9B
- + Successfully incorporates all elements: tv anchor desk, dogs, and a hockey arena setting.
- + Captures the user's likeness well in a caricature art style.
- + Highly creative and humorous interpretation with many fine details.
- − Text on the screens and boards is nonsensical.
- − The microphone and pen held by the character have slightly confused geometry.
Z-Image Turbo
- + Maintains the high photographic quality of the source image.
- − Completely failed the edit instruction to create a caricature.
- − Did not incorporate the hockey or tv anchor themes.
- − The only addition is a small, blurry dog in the background.
Verdict: FLUX.2 [klein] 9B followed the complex instructions perfectly, creating a vibrant caricature that combined tv broadcasting, hockey, and dogs while maintaining a recognizable likeness of the woman in the source image. Z-Image Turbo almost entirely ignored the prompt, failing to change the art style or add anything other than a barely visible dog in the background.
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.2 [klein] 9B
- + Exceptional lighting effects with prominent god rays and golden hour ambiance
- + High level of detail in the fur textures and butterfly designs
- + Vibrant and dense wildflower meadow composition
- − Missed one of the four requested animals (the baby bunny is absent)
- − The kitten's pose and expression appear slightly distorted and less 'photorealistic'
Z-Image Turbo
- + Successfully included all four requested animals: puppy, kitten, fox, and bunny
- + Great interaction between the animals, particularly the puppy's paw on the rabbit
- + Natural and adorable facial expressions on all creatures
- − Lighting is less dramatic compared to the 'god rays' requested in the prompt
- − Slight lack of sharpness in the background elements compared to Model A
Verdict: While FLUX.2 [klein] 9B produced a more visually striking image with superior lighting and texture detail, it failed the prompt adherence test by omitting the baby bunny. Z-Image Turbo followed the complex multi-subject prompt perfectly, including all four animals with a natural and wholesome layout, making it the more successful interpretation despite slightly softer lighting.
Studio Ghibli Anime Style
Editing“Transform this photo into a Studio Ghibli–inspired illustration. Use soft pastel colors, hand-painted textures, gentle lighting, dreamy backgrounds, and a warm, nostalgic mood”
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellently captures the requested Studio Ghibli art style with watercolor textures and soft line art.
- + Precisely replicates the 'distracted boyfriend' meme composition while adapting it to the animation style.
- + Successfully transforms the background into a dreamy, Ghibli-esque landscape as requested.
- − The facial expression of the partner on the right lost its original 'shocked' intensity in favor of a pleasant look.
Z-Image Turbo
- + Perfectly preserves the source image composition and the exact facial features of the original people.
- + Retains the high-resolution realism of the source photo.
- − Completely failed to apply the requested edit of a Studio Ghibli illustration style.
- − Image remains a standard photograph with no visible artistic textures or pastel colors.
Verdict: FLUX.2 [klein] 9B successfully followed the creative brief, transforming the meme into a beautiful anime-style illustration with hand-painted textures and a nostalgic feel. Z-Image Turbo essentially ignored the stylistic instructions, providing a near-identical copy of the source image with almost no perceptible changes.
Golden Hour Stroll
Image Editing“Add dynamic motion to this photo: make hair blow in the wind, add leaves flying, energetic and lively feel.”
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent adherence to the 'hair blowing' instruction with naturalistic flow
- + Abundant and dynamic flying leaves that create a sense of depth
- + Very high source preservation, maintaining the woman's face and dog's appearance perfectly
- − Some leaves appear to be floating in front of the dog's fur without proper interaction
Z-Image Turbo
- + Successfully adds a subtle wind effect to the hair and some scattered leaves
- + Maintains a clean aesthetic with logical lighting on the new elements
- − Less 'energetic' feel compared to model A
- − Noticeable changes to the woman's facial features and the bridge in the background
- − The dog's leash has been partially removed/disconnected from the woman's hand
Verdict: FLUX.2 [klein] 9B followed the prompt more effectively, delivering a much higher volume of motion-inducing elements like leaves and more dramatic hair movement while perfectly preserving the identity of the human and animal subjects. Z-Image Turbo struggled with source preservation, altering the woman's face and accidentally removing the leash, while providing a much more timid interpretation of 'dynamic motion'.
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.2 [klein] 9B
- + Excellent typography rendering with the requested accent mark.
- + Professional vector emblem style with a cohesive circular layout.
- + Successfully included the 'Est. 1720' banner as requested.
- − The steam trail is a bit chunky compared to the rest of the minimal design.
- − The background texture is somewhat blotchy.
Z-Image Turbo
- + Clean, minimalist silhouette with a modern flat design feel.
- + Accurate text spelling and nice negative space within the cloche.
- + Good implementation of the warm brown and cream color scheme.
- − Failed to include the requested banner for the 'Est. 1720' text.
- − The steam icon is very small and lacks the 'vintage' character of the first model.
Verdict: FLUX.2 [klein] 9B followed the complex prompt requirements more closely, specifically by including the 'Est. 1720' banner and creating a more traditional vector emblem. While Z-Image Turbo produced a clean and viable logo, it missed the banner element and had a more generic icon style.
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.2 [klein] 9B
- + Stronger composition with a logical visual flow and pathing.
- + Better adherence to the requested NASA-inspired color palette.
- + Includes more distinct steps as requested in the prompt.
- − Numerous spelling errors in labels such as 'AOILLO', 'EARDHT', and 'LANDINING'.
- − The logical flow of the icons is somewhat circular and confusing to follow.
Z-Image Turbo
- + Cleaner vector aesthetic with very crisp lines.
- + More accurate iconography for the Lunar Module.
- + Excellent use of negative space for a modern look.
- − Missed several prompt instructions including the specific steps and color palette.
- − Spelling errors in the main header and labels.
- − Very poor composition with floating text and disjointed elements.
Verdict: FLUX.2 [klein] 9B followed the structural instructions of the prompt much better, providing a full narrative flow from launch to landing, even though it struggled significantly with spelling. Z-Image Turbo produced cleaner individual icons but failed to include all the requested steps and had a very poor, unorganized layout. FLUX.2 is the winner for better prompt adherence and a more complete infographic feel.
Explore each model
Tongyi-MAI's 6-billion parameter distilled text-to-image model optimized for speed, achieving high-quality generation in 8 steps or fewer with support for bilingual text rendering