Black Forest Labs' distilled 9 billion parameter image generation model with sub-second inference and multi-reference support
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [klein] 9B
20.6
arena score
#13 of 32 in Image Editing
Skill signature
· Image Editing
Imagen 4.0 Ultra Generate 001
21.9
arena score
#33 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [klein] 9B
0%
win rate
Ties
0%
Imagen 4.0 Ultra Generate 001
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent depiction of a glass container having physical thickness and realistic caustics on the wood.
- + High level of photographic realism in the wooden table texture and plant background.
- − The glass container resembles a vase or heavy jar rather than a geometric cube.
- − The perspective on the red book is slightly skewed relative to the top of the glass.
Imagen 4.0 Ultra Generate 001
- + Perfectly depicts a solid geometric glass cube as requested.
- + Impressive text rendering on the book's spine and sharp lighting highlights.
- + Stronger adherence to the prompt regarding the plant being visible 'through the glass'.
- − The blue sphere has a strange double-image or motion blur artifact inside the cube.
- − The plant's color saturation is slightly less natural than in the other image.
Verdict: Both models followed the complex spatial instructions well. FLUX.2 [klein] 9B produced a more naturally lit and photorealistic scene, but the 'cube' is rounded and looks more like a jar. Imagen 4.0 Ultra Generate 001 captured the geometric cube shape and specific details like the book's title much better, making it the winner for prompt adherence despite a small artifact behind the blue sphere.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent skin texture and hyper-realistic facial details
- + Accurate depiction of a 50mm lens depth of field with cinematic lighting
- + Complex reflections on the wet pavement that feel grounded in the scene
- − Lack of noticeable motion blur from the passing car as requested
Imagen 4.0 Ultra Generate 001
- + Successfully captured the 'imperfect framing' and candid street photography vibe
- + Effective shallow depth of field and plausible wet environment textures
- + Good interaction between the hands and the bicycle chain mechanism
- − Lack of required motion blur on the background vehicle
- − The bicycle geometry is slightly warped, specifically the handlebar and frame intersection against the wall
Verdict: FLUX.2 [klein] 9B delivers a superior level of fine detail in the subject's skin and clothing, alongside more sophisticated lighting and reflections that match the 'cinematic' prompt. While Imagen 4.0 Ultra captures the 'imperfect framing' of a candid shot better, it suffers from minor geometric inconsistencies in the bicycle and lacks the textural depth found in the FLUX output.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent skin texture with realistic pores and fine details
- + The lighting feels more integrated with the scene
- + Superior rendering of leather straps with worn edges and micro-textures
- − The braids are less elaborate and detailed compared to the other model
Imagen 4.0 Ultra Generate 001
- + Highly detailed hair braiding with consistent bead placement
- + Strong representation of the engraved plate armor
- + Good inclusion of all requested elements including the torch and sparks
- − The skin texture looks slightly more digital/artificial
- − The facial expression is a bit stiff
Verdict: Both models followed the prompt exceptionally well, capturing the battle-worn aesthetic and ornate armor. FLUX.2 [klein] 9B is the winner due to its superior photographic realism in the skin, eyes, and leather textures, whereas Imagen 4.0 Ultra Generate 001, while excellent in its braiding detail, has a slightly more artificial character render.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Stronger prompt adherence regarding the section structure (appetizers, pizza, mains).
- + Effective use of bold sans-serif fonts for a professional brand look.
- + Good color accents that provide visual interest without clutter.
- − The food photos are repetitive, showing four pizzas under different category headings.
- − Significant text corruption and 'gibberish' in the main menu items.
- − The layout becomes cramped in the bottom half.
Imagen 4.0 Ultra Generate 001
- + Excellent grid layout with a wide variety of distinct food photos.
- + Much higher image clarity and cleaner minimalist aesthetic.
- + Better spacing and whitespace management for a modern look.
- − Missed the request for a specific 'Mains' text section below the grid.
- − Section headers are slightly misaligned with the grid columns below.
- − Text is still largely nonsensical, though slightly more legible than the competitor.
Verdict: Imagen 4.0 Ultra is the preferred choice due to its superior visual quality and the diversity of its food photography, which makes it look like a real menu. While FLUX.2 followed the structural instructions for section titles more closely, it failed significantly by using identical pizza images for every category, whereas Imagen 4.0 Ultra provided a high-quality, professional grid layout.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent typography with a professional, cohesive 'fiery' texture applied to all text components.
- + Highly photographic lettuce and meat texture that looks very appetizing.
- + Dynamic composition with diagonal movement and high-energy particle effects.
- − The burger is not truly 'exploded'; most components are still touching or only slightly separated.
- − The sauce drops look somewhat like plastic or jelly rather than liquid.
Imagen 4.0 Ultra Generate 001
- + Perfect adherence to the 'exploded' instruction with clear vertical separation between all ingredients.
- + Excellent rendering of the 'fiery' effect on the text and the starburst element.
- + Very clean, symmetrical layout that works well for a commercial advertisement.
- − The background swirl effect creates a slightly more 'magical' than 'photorealistic' feel compared to A.
- − The tomatoes look a bit stylized or CG compared to the realism of the patties in the other image.
Verdict: Both models followed the prompt exceptionally well, particularly regarding the complex text requirements. Imagen 4.0 Ultra Generate 001 is the overall winner because it better captures the 'exploded' burger concept with distinct separation between layers, whereas FLUX.2 [klein] 9B kept the burger mostly intact.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent chalk texture with realistic smudges and varying pressure.
- + Highly accurate handwritten appearance with natural variations in letter size.
- + Perfectly captures the requested 'elegant cursive' for the title.
- − Contains a minor spelling error ('fress' instead of 'fresh').
Imagen 4.0 Ultra Generate 001
- + Perfect spelling across all text on the board.
- + Clean and legible composition.
- + Good contrast against the lighting at the top of the frame.
- − The text looks a bit more like a digital font than natural hand-drawn chalk.
- − Fails the 'elegant cursive' requirement for the title, using a print-leaning script instead.
- − Chalk texture on the letters is very uniform and lacks the 'dusty' realism of the other image.
Verdict: FLUX.2 [klein] 9B produces an incredibly realistic chalk board with authentic textures and beautiful handwritten styles that feel human-made, though it suffers from one small typo. Imagen 4.0 Ultra Generate 001 is functionally perfect in its spelling but lacks the organic character and artistic cursive requested in the prompt, resulting in a slightly more 'digital' look. FLUX.2 is the winner for its superior texture and adherence to the stylistic handwriting requests.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Vibrant and cinematic background with detailed nebulae and planets.
- + High level of texture detail on the horse's coat and space suit.
- + Dynamic composition that feels more surreal and expansive.
- − Fails the negative constraint, depicting an astronaut riding a horse.
- − Includes distracting, nonsensical AI-generated text at the bottom.
- − Anatomy of the horse's back legs and hoof placement is slightly awkward.
Imagen 4.0 Ultra Generate 001
- + Creative inclusion of a space helmet and glowing hooves for the horse.
- + Very clean, high-resolution rendering with professional lighting.
- + More cohesive theme with the horse standing on a cratered asteroid.
- − Completely ignores the specific 'horse on top' prompt instruction.
- − Relatively safe and standard composition for an 'astronaut on horse' prompt.
- − Large white letterboxing borders are present in the output.
Verdict: Both models failed the specific prompt instruction to place the horse on top of the astronaut, instead providing the common 'astronaut on a horse' trope. FLUX.2 [klein] 9B offers a more cinematic and detailed cosmic environment but is marred by gibberish text, while Imagen 4.0 Ultra Generate 001 provides better costume details for the horse and cleaner overall image quality despite the smaller canvas size.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent photorealistic lighting and textures inside the vehicle.
- + Perfectly captures the 'bored' expression of the passenger as requested.
- + Authentic wide-angle composition from the perspective of the dashboard.
- − Nonsense text on the taxi cap ('NEW TALA').
- − Technical anatomy issue where the capybara appears to have four legs/paws in the front area.
Imagen 4.0 Ultra Generate 001
- + Sharp, clear rendering of the capybara's fur and clothing.
- + Correct text rendering on the driver cap.
- + Dynamic lighting that highlights the city bokeh effectively.
- − Compositional error: the passenger is in the front passenger seat, not the back seat as requested.
- − The image has a slightly more stylized/digital look rather than pure photorealism.
- − The passenger's expression looks stressed rather than bored.
Verdict: FLUX.2 [klein] 9B followed the spatial instructions much better by placing the passenger in the back seat and capturing the specific 'monotone' mood of a New York night. While Imagen 4.0 Ultra provided clearer text and sharper detail on the capybara, it failed the core requirement of placing the businesswoman in the back seat and gave her an anxious expression rather than a bored one.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Perfect text rendering for all lines including the specific date and location.
- + Superior atmospheric lighting and more cohesive 'dark parchment' texture.
- + Traditional gothic layout with well-integrated thorns and cobwebs in the border.
- − The moon is slightly obscured by the text compared to the other model.
Imagen 4.0 Ultra Generate 001
- + Clean, vibrant colors and very clear scroll banner implementation.
- + Dynamic use of thorns that break the border frame for a 3D effect.
- + Accurate text rendering for all requested details.
- − The art style leans more towards digital illustration than 'vintage gothic' or 'cinematic'.
- − The background feels a bit more generic and less atmospheric than the first option.
Verdict: Both models followed the prompt instructions perfectly, including the complex text requirements. FLUX.2 [klein] 9B is the winner because its aesthetic more closely matches the 'vintage gothic' and 'cinematic' keywords, whereas Imagen 4.0 Ultra feels more like a modern vector illustration.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent 3D miniature 'toy' aesthetic with soft lighting.
- + Perfect text rendering for both 'JAPAN' and 'SUSHI'.
- + Strong isometric perspective on a wooden diorama base.
- − The flag depicted is not the Japanese flag.
- − Visible clipping between the salmon nigiri and the maki rolls.
Imagen 4.0 Ultra Generate 001
- + Correctly identifies and renders the Japanese flag icon.
- + Higher variety of sushi types adding visual interest.
- + Clean, professional layouts and more polished materials.
- − The text 'JAPAN' is not perfectly centered above 'SUSHI' as requested (flag is to the right).
- − The 'miniature' cartoon feel is slightly less pronounced compared to the competitor.
Verdict: Both models followed the prompt well, but Imagen 4.0 Ultra is the winner for correctly identifying the Japanese flag, whereas FLUX.2 rendered a generic red and black horizontal tricolor. While FLUX.2 had a slightly more charming 'cartoon' texture, Imagen 4.0 provided a more comprehensive and accurate scene with better-defined assets.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent photographic lighting with realistic backlighting and god rays.
- + Ultra-detailed fur texture that appears soft and natural.
- + Realistic bokeh and depth of field in the floral foreground.
- − Missing the baby bunny requested in the prompt.
- − Anatomy of the kitten is slightly awkward with some merged fur/limbs.
Imagen 4.0 Ultra Generate 001
- + Includes all four requested animals: puppy, kitten, fox, and bunny.
- + Bright, vibrant color palette that captures the 'joyful wholesome vibe'.
- + Clear interaction between the animals and the butterflies.
- − Less 'photorealistic' and more illustrative/CG in appearance.
- − Dew sparkles look like flat white circles rather than natural light reflections.
- − Some anatomical issues with the fox's floating back leg.
Verdict: Imagen 4.0 Ultra follows the prompt more comprehensively by including the baby bunny which FLUX.2 [klein] 9B omitted. However, FLUX.2 [klein] 9B produces a significantly more high-end photographic image with superior light handling and texture, while Imagen feels more like an polished digital illustration.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Excellent typography with clean, thick lines and correct accent marks.
- + Strong emblem composition using a circular frame and ribbon.
- + Deep, rich color palette that fits the 'warm brown' prompt perfectly.
- − The steam element looks slightly like a flame or a decorative flourish rather than vapor.
- − The 'Est. 1720' text is slightly off-center within the ribbon.
Imagen 4.0 Ultra Generate 001
- + Elegant minimalist aesthetic with great use of negative space.
- + Accurate representation of steam rising from the cloche.
- + Clean cross-hatching detail on the cloche adds to the vintage feel.
- − The 'EST. 1720' text is poorly integrated, appearing flat and off-perspective relative to the ribbon's curves.
- − The composition feels a bit disjointed with the line separating the logo elements.
Verdict: Both models followed the prompt closely, creating high-quality vector-style logos. FLUX.2 [klein] 9B is the winner because its typography and emblem construction feel like a cohesive, professional brand identity, whereas Imagen 4.0 Ultra Generate 001 produced a ribbon with poorly warped text that breaks the vintage aesthetic.
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 [klein] 9B
Imagen 4.0 Ultra Generate 001
AI Judge Analysis
FLUX.2 [klein] 9B
- + Strong adherence to the requested icons like Saturn V and Lunar Module.
- + Clear, large, and legible typography despite some spelling errors.
- + Clean flat-vector style with a layout that is easy to follow.
Imagen 4.0 Ultra Generate 001
- + Excellent adherence to the color palette (navy, white, muted red).
- + Sophisticated professional infographic layout with structured columns.
- + Sophisticated text rendering for the main titles.
- − Failed to provide the specific icons requested for each step.
- − The 'steps' are abstract and do not clearly represent the mission progression as described.
- − Nonsensical placeholder text for all descriptions.
Verdict: FLUX.2 [klein] 9B followed the content instructions much better, providing the specific icons and sequence requested (Launch, Earth, Moon, Descent). While Imagen 4.0 Ultra Generate 001 produced a much more realistic 'infographic' look and better followed the color palette, its content was largely generic or irrelevant to the specific Apollo 11 steps requested.
Next steps
Explore each model
Google's Imagen 4.0 Ultra model offering the highest fidelity and resolution for professional-grade image generation