Black Forest Labs' open-weights image generation model with frontier performance, available for non-commercial local deployment
Settled by community votes across 13 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev]
24.3
arena score
#19 of 62 in Text-to-Image
Skill signature
· Text-to-Image
Imagen 4.0 Generate 001
17.0
arena score
#55 of 62 in Text-to-Image
Vote tally
Where the votes landed
FLUX.2 [dev]
0%
win rate
Ties
0%
Imagen 4.0 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 [dev]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to spatial relationships, specifically placing the plant behind the cube.
- + Realistic glass transparency and reflections of both the sphere and the plant.
- + Accurate lighting that matches the 'soft window light from the left' instruction.
Imagen 4.0 Generate 001
- + Sharp focus on the textures of the red book.
- + Clean, modern aesthetic with high contrast.
- − The sphere appears to be floating mid-air inside the cube without support.
- − The plant is not visible through the glass as requested, as the cube surfaces are too reflective/mirror-like.
- − Perspective distortion makes the book look significantly larger than the cube's footprint.
Verdict: FLUX.2 [dev] followed the complex spatial instructions perfectly, accurately depicting the plant behind the transparent cube and the sphere resting on the bottom surface. Imagen 4.0 Generate 001 struggled with the transparency physics, resulting in a cube that looks more like a mirror and a sphere that is inexplicably floating.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent execution of motion blur from passing cars
- + Extremely realistic and natural skin texture with visible age spots and dirt
- + Highly convincing rain effects and surface moisture on the jacket and pavement
- − The hands have some slight anatomical confusion where they grip the handlebars
Imagen 4.0 Generate 001
- + Strong composition with vibrant red bike and colorful city lights
- + Good interaction with a specific tool (screwdriver) and the bike chain
- − Failed to include motion blur from passing cars
- − The image feels more like a posed portrait rather than the requested 'candid street photo'
- − Anatomical issues with the subject's right hand holding the tool
Verdict: FLUX.2 [dev] followed the prompt more accurately, particularly regarding the 'candid' feel and the inclusion of motion blur in the background. While Imagen 4.0 created a vibrant scene, it looked more like a staged photo and missed the specific request for passing vehicle motion blur, whereas FLUX.2's skin details and ambient rain effects felt more authentic.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent adherence to the 'beads in braids' prompt with colorful, variety of beads.
- + High realism in facial textures, scars, and lifelike eye details.
- + Superior representation of battle-worn dirt and grime on the skin.
- − The torch light is slightly cut off at the edge of the frame.
Imagen 4.0 Generate 001
- + Ornate engraving on the plate armor is exceptionally sharp and intricate.
- + Dramatic lighting with strong warm and cool contrast.
- + The leather strap texture is very tactile and clear.
- − The 'beads' in the hair look more like metallic cylinders or hardware rather than decorative beads.
- − The skin looks a bit overly smooth and plastic-like compared to Model A.
- − The 'braids' are stiff and less natural in appearance.
Verdict: FLUX.2 [dev] followed the specific details of the prompt much better, particularly the beads and the battle-worn skin textures, which feel authentic and gritty. While Imagen 4.0 has impressive armor engravings and lighting, the hair details feel more like futuristic tech than a medieval paladin, and the character lack the realism found in the competitor's output.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography with clean, bold sans-serif fonts
- + High-quality, realistic food photography
- + Effective use of vibrant color accents for section headers
- − The 'PIZZAU' header contains a misspelling
- − Text alignment becomes slightly messy in the right column
Imagen 4.0 Generate 001
- + Perfectly executed grid layout representing the 'modern minimalist' request
- + High-end plating and lighting in food photos
- + Clean and legible font choices for menu items
- − Pricing format is simplified to integers, lacking currency symbols or decimals
- − Geometric decorative elements are a bit generic
Verdict: FLUX.2 [dev] produces a very traditional and professional-looking long-form menu with excellent photo quality, but suffers from a noticeable typo in one of the main headers. Imagen 4.0 captures the 'modern minimalist' and 'grid' prompt much more effectively with a creative square layout, better source preservation of the white space, and slightly cleaner overall composition for a casual dining vibe.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography with a cohesive flaming neon aesthetic
- + Clean, high-quality rendering of the starburst effect
- + Photorealistic food textures, especially the sear on the patty
- − The 'exploded' effect is less dynamic than the other model, with components grouped closely
- − The background is slightly more generic in its composition
Imagen 4.0 Generate 001
- + Very dynamic composition with a strong sense of motion and spread
- + Includes many extra details like onions and pickles not explicitly in the prompt
- + Good color contrast between the dark background and the vibrant ingredients
- − Text rendering is slightly weaker, with less consistent glowing effects
- − The price tag uses a comma instead of a period (6,99), which may not fit all regions
- − Text placement overlaps some of the ingredients, making it feel cluttered
Verdict: FLUX.2 [dev] produces a more polished and professional advertisement with superior typography and lighting that feels integrated into the scene. While Imagen 4.0 Generate 001 offers a more energetic 'exploded' layout, its text placement and punctuation error on the price make it less effective as a final marketing asset.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent chalk texture including realistic smudges and varying stroke weight
- + Accurate spelling and completion of the logical end of the prompt's cut-off text
- + Authentic cozy café background that adds depth and context
Imagen 4.0 Generate 001
- + Clean layout centered within a wooden frame
- + Legible handwriting style
- − Included meta-instructions from the prompt as actual text on the board
- − Numerous spelling errors and nonsensical sentences
- − Texture looks like a digital font rather than natural chalk
Verdict: FLUX.2 [dev] significantly outperformed Imagen 4.0 by understanding the intent of the prompt and rendering a realistic chalkboard with excellent texture and logical completion of the text. Imagen 4.0 failed by literalizing the prompt's descriptive instructions (e.g., writing the words 'Footer' and 'Tittle Menu' on the board) and suffered from poor spelling and a lack of authentic chalk texture.
The Reversed Rodeo
Text-to-Image“Horse riding astronaut in space — horse on top, not vice versa. Surreal, highly detailed, cinematic.”
FLUX.2 [dev]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Realistic cinematic lighting and textures.
- + Clear, high-resolution rendering of the space background and Earth.
- + Natural-looking horse anatomy and fur.
- − Completely failed the negative constraint to have the horse on top of the astronaut.
Imagen 4.0 Generate 001
- + Stronger surreal aesthetic with galaxy patterns in the horse's fur.
- + Excellent composition with rainbow nebulae and cosmic dust.
- + High level of detail in the spacesuit design.
- − Completely failed the negative constraint to have the horse on top of the astronaut.
Verdict: Both FLUX.2 [dev] and Imagen 4.0 failed to follow the specific spatial instruction to place the horse on top of the astronaut, instead defaulting to the common trope of an astronaut riding a horse. Imagen 4.0 is the winner as it delivered a more creative and visually stunning 'surreal' interpretation of the scene, whereas FLUX.2 [dev] provided a more literal, standard cinematic look.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent photorealism with shallow depth of field.
- + Great texture on the capybara's fur and the passenger's coat.
- + Realistic framing from outside the windshield looking in.
- − The capybara's paws look more like human hands with animal skin.
- − The passenger appears to be in the front seat or very close behind the driver due to perspective.
Imagen 4.0 Generate 001
- + Perfect composition clearly showing the back-seat passenger and front-seat driver.
- + The capybara's paws are rendered more realistically as animal feet.
- + Include specific requested details like the 'Taxi' text on the hat.
- − Slightly more 'digital' or CGI look compared to Model A.
- − The seatbelt on the passenger is positioned slightly awkwardly.
Verdict: Both models followed the prompt exceptionally well. FLUX.2 [dev] produces a more convincing photographic look with better lighting and textures, but Imagen 4.0 captures the layout and spatial relationship between the driver and passenger more accurately.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography with perfect adherence to the requested text and formatting.
- + High visual quality with realistic 'dark parchment' textures and moody cinematic lighting.
- + Seamless integration of the thorny border and cobwebs within the composition.
- − The jack-o-lantern is quite large, slightly overwhelming the detail text below it.
Imagen 4.0 Generate 001
- + Strong gothic font choice for the main heading.
- + Clean layout for the event details at the bottom using vertical separators.
- + Atmospheric use of the moon and silhouettes for depth.
- − Includes a strange, vertical rolled scroll artifact on the left side that clutters the image.
- − The banner text is slightly distorted on the left edge.
- − The jack-o-lantern is a bit small compared to the scale of the trees.
Verdict: FLUX.2 [dev] followed the prompt more effectively, producing a cohesive and polished invitation with perfect text rendering and superior textures. Imagen 4.0 Generate 001 produced a visually interesting scene but was hindered by a strange vertical scroll artifact and less defined parchment textures. FLUX.2 [dev] is the clear winner for its professional layout and adherence to the 'dark parchment' aesthetic.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent typography adherence, perfectly rendering the requested text and flag icon.
- + Strict adherence to the isometric perspective and composition prompts.
- + Clean, professional miniature 3D aesthetic with a solid light blue background as requested.
- − The textures are a bit simplified/matte compared to the 'realistic PBR' request.
Imagen 4.0 Generate 001
- + High visual realism and impressive textures on the fish and roe.
- + Effective use of lighting to create depth and dimension.
- − Completely failed to include the requested text ('JAPAN', 'SUSHI') and flag icon.
- − Background is grey/white instead of the requested light blue.
- − Angle is a low-angle perspective rather than the requested 45° top-down isometric view.
Verdict: FLUX.2 [dev] followed every instruction in the prompt, including complex layout requirements like specific typography, icon placement, and isometric perspective. While Imagen 4.0 Generate 001 produced a more photorealistic image of food, it ignored the text, background color, and specific camera angle requirements, making FLUX.2 [dev] the clear winner for prompt adherence.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent photorealistic lighting and backlighting effects.
- + High degree of anatomical realism in the animals' fur and features.
- + Beautifully soft cinematic atmosphere with realistic depth of field.
- − Included two bunnies instead of one.
- − The animals are sitting rather than 'tumbling and chasing' as requested.
Imagen 4.0 Generate 001
- + Captures the 'tumbling' and active movement much better than the competitor.
- + Exactly matches the count of each animal requested.
- + Very vibrant colors and clear dew drop details on the grass.
- − Has a more 'illustrated' or digital art feel rather than hyper-photorealistic.
- − The fox's anatomy and the cat's belly look slightly artificial.
Verdict: FLUX.2 [dev] produces a significantly more realistic and cinematic image with superior lighting and texture, though it failed the exact count of animals and produced a static pose. Imagen 4.0 followed the action and count instructions much more closely, but the resulting image lacks the high-end photographic quality of FLUX.2, feeling more like a high-quality greeting card 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 [dev]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent woodcut/vintage aesthetic with high-quality grain texture.
- + Flawless typography including the accent on the 'E'.
- + Strong banner integration that feels cohesive with the emblem style.
- − The banner layout is a bit busy for a truly 'minimalist' request.
Imagen 4.0 Generate 001
- + Successfully achieves a minimalist, clean vector look.
- + Clear, legible typography and well-defined cloche illustration.
- + Accurate adherence to the color palette.
- − The 'Est. 1720' banner is quite small and lacks the 'classic' flair of Model A.
- − Lacks the requested 'subtle texture' on the background.
Verdict: FLUX.2 [dev] captures the vintage, classic aesthetic much more effectively with superior typography and texture that creates a premium feel. While Imagen 4.0 is cleaner and more minimalist, it feels a bit generic compared to the rich, woodcut style and elegant banner design provided by FLUX.2. Both models handled all text elements perfectly.
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]
Imagen 4.0 Generate 001
AI Judge Analysis
FLUX.2 [dev]
- + Excellent vector iconography with clear detail in the lunar module and Saturn V.
- + Stronger adherence to the NASA-inspired color palette.
- + Includes names of the astronauts in a creative logo format.
- − The layout is cluttered and the steps are not in a logical chronological order.
- − Significant text artifacts and misspellings for several labels.
Imagen 4.0 Generate 001
- + Layout captures the infographic feel better with a clear vertical flow.
- + Excellent text rendering with no spelling errors in the labels.
- + Clean, minimalist vector aesthetic that feels more professional.
- − Missed the final two steps of the prompt (Descent and Landing).
- − The icons for Launch and Translunar are conceptually confusing or repetitive.
Verdict: FLUX.2 [dev] includes more of the requested steps and detailed iconography, but fails significantly on layout logic and text accuracy. Imagen 4.0 Generate 001 provides a much cleaner, more 'finished' infographic look with perfect text, though it truncated the requested list of steps. Imagen 4.0 is the preferred choice for a usable graphic despite the missing steps because its composition and typography are superior.
Next steps
Explore each model
Google's latest Imagen 4.0 text-to-image generation model with significantly better text rendering and overall image quality