Black Forest Labs' open-weights image generation model with frontier performance, available for non-commercial local deployment
Settled by community votes across 15 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev]
#18 of 62 in Text-to-Image
OmniGen v2
#57 of 62 in Text-to-Image
Where the votes landed
FLUX.2 [dev]
0%
win rate
Ties
0%
OmniGen v2
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 [dev]
- + Excellent surface textures with realistic dust and reflections on the glass
- + Superior lighting consistency with logical shadows on the wooden table
- + Perfect adherence to the plant visibility through the glass.
- − The plant is slightly more overexposed by the window light than expected.
OmniGen v2
- + Clean, vibrant colors with a very smooth blue sphere
- + Good spatial arrangement of the requested objects
- − The glass cube lacks a top pane, making the book appear to float improperly
- − The reflection of the sphere on the bottom of the cube is missing or physically incorrect
- − The lighting on the cube and sphere is somewhat flat compared to the background.
Verdict: FLUX.2 [dev] produced a much more realistic and physically coherent image, with the glass cube feeling like a solid object that correctly refracts the plant behind it. OmniGen v2 struggled with the physics of the glass, notably the missing top surface and inconsistent reflections, resulting in a more 'rendered' and less photographic look.
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 [dev]
- + Excellent adherence to the 'candid' and 'imperfect framing' requirements.
- + Highly realistic textures on the skin, wet fabric, and mechanical parts of the bike.
- + Effective use of motion blur on passing vehicles to create a sense of scale and movement.
- − The hands have structural anatomical issues with finger count and overlapping layers.
OmniGen v2
- + Strong composition with clear reflections and a clean red bicycle.
- + Good color contrast and mood with the light rain streaks.
- − Appears overly staged and lacks the requested 'candid' street photography aesthetic.
- − The cars in the background are static and lack the requested motion blur.
- − The man's stance and the bicycle's proportions look slightly awkward and digital.
Verdict: FLUX.2 [dev] significantly outperforms OmniGen v2 in technical adherence to the photography-specific prompts, successfully capturing the 'candid' look and the motion blur of passing cars. While FLUX.2 struggles with hand anatomy, its skin textures and overall realism align much better with the 'no stylization' request compared to the more illustrative and plastic feel of OmniGen v2.
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 [dev]
- + Excellent adherence to the 'battle-worn' description with realistic scars and grime.
- + Superior metal engraving detail and complex light reflections.
- + Highly detailed texture on leather straps and cloth elements.
OmniGen v2
- + Strong composition with a clear focal point and shallow depth of field.
- + Clean skin textures and expressive eyes.
- − Fails to capture the 'battle-worn' aesthetic, appearing too clean and polished.
- − Armor engraving is less intricate than requested.
- − Beads are less organically integrated into the hairstyle.
Verdict: FLUX.2 [dev] successfully captured all the nuances of the prompt, particularly the 'battle-worn' characteristics and the intricate textures of the armor and leather. In contrast, OmniGen v2 produced a much cleaner, more stylized portait that missed the grit and detail requested in the engraving and scarring.
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 [dev]
- + Excellent typography rendering with almost legible text
- + Clean, professional layout that feels like a real restaurant menu
- + High-quality, realistic food photography photography with consistent lighting
- − Text is still mostly gibberish upon close inspection
- − Layout is a bit crowded with many sections overlapping
OmniGen v2
- + Strong minimalist aesthetic with bold color blocking
- + Unique grid layout with centered food photography
- − Severe spelling errors in headers (e.g., 'APPTETIZES', 'RESTAURATED MENTS')
- − Body text is completely unreadable greeking/gibberish lines
- − Less professional arrangement for a functional menu
Verdict: FLUX.2 [dev] produces a much more professional and realistic menu design that adheres to the 'modern minimalist' and 'casual dining' requirements. While OmniGen v2 has creative color blocking, its significant spelling errors and unreadable body text make it less usable as a graphic design asset compared to the cleaner execution of FLUX.2 [dev].
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 [dev]
- + Excellent adherence to the 'exploded' burger request with clearly suspended components.
- + Flawless rendering of all requested text and the fiery starburst element.
- + Very high photorealistic detail in the food textures, especially the patty and bun.
- − The sauce droplets look a bit frozen rather than having a dynamic motion blur.
- − The composition feels slightly crowded at the top.
OmniGen v2
- + Clean, graphic-style composition with a bright, eye-catching central focal point.
- + Integrated a double patty which adds to the 'magic' or deluxe feel of the ad.
- − Completely failed the 'exploded' and 'suspended in mid-air' instruction, showing a static burger.
- − Text rendering is poor, with 'LIMITED TIME ONLY' missing letters and the currency symbol omitted from the price.
- − Visual style is more cartoonish/3D-rendered rather than the requested photorealistic detail.
Verdict: FLUX.2 [dev] followed every instruction perfectly, delivering a high-quality exploded burger with accurate text and a realistic fiery atmosphere. OmniGen v2 failed the primary structural request for an exploded view and struggled with text legibility and accuracy, resulting in a generic static advertisement.
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 [dev]
- + Excellent text rendering with near-perfect spelling across the entire board.
- + Highly realistic chalk texture with smudges and natural handwriting variations.
- + Matches the requested cozy cafe atmosphere with depth of field in the background.
- − The date is in a print-style sans-serif rather than the requested elegant cursive.
OmniGen v2
- + Successfully captured the requested date and price segments.
- − Numerous spelling errors including 'Specals', 'Lemont', and 'Chocola'.
- − The text layout is cluttered and overlapping, especially in the middle section.
- − Lacks the realistic texture and atmosphere of a real café setting.
Verdict: FLUX.2 [dev] significantly outperforms OmniGen v2 by providing legible, accurately spelled text and a much more realistic chalk aesthetic. While OmniGen v2 struggled with basic spelling and layout consistency, FLUX.2 [dev] successfully rendered several long phrases and maintained a high level of visual quality throughout.
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 [dev]
- + Excellent cinematic lighting and atmosphere.
- + High level of detail on the space suit and horse textures.
- + Realistic integration of the character within the lunar/orbital environment.
- − Failed the negative constraint; the astronaut is riding the horse.
OmniGen v2
- + Substance of the astronaut and horse is clear and vibrant.
- + Correctly interprets the space theme with planets/moons.
- − Failed the negative constraint; the astronaut is riding the horse.
- − The composition is much simpler and looks like a basic clip-art style.
- − Anatomical issues with the horse's legs and the floating reins.
Verdict: Both models failed the specific spatial logic of the prompt ('horse on top, not vice versa'), which was a test of prompt adherence for surrealism. FLUX.2 [dev] is the superior image because it achieves a high-quality cinematic aesthetic with realistic lighting and textures, whereas OmniGen v2 produced a flat, lower-quality illustration with anatomical errors.
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 [dev]
- + Excellent adherence to the capybara's physical traits, including realistic fur and paws on the wheel.
- + Accurate positioning of the passenger in the back seat as requested.
- + Realistic cinematic lighting and depth of field that conveys a New York night atmosphere.
- − The passenger's hands and face are slightly soft in terms of focus and detail compared to the foreground.
OmniGen v2
- + High image sharpness and clean colors.
- + The capybara's clothing looks like a high-quality professional jacket.
- − Failed the spatial request by putting the passenger in the front seat holding the steering wheel.
- − The capybara has human hands, which is a significant anatomical error.
- − The perspective is warped, making the capybara appear to be in the passenger seat while the woman drives.
Verdict: FLUX.2 [dev] significantly outperformed OmniGen v2 by correctly following the spatial and anatomical instructions. While FLUX.2 [dev] accurately depicted a capybara with paws driving and a passenger in the back, OmniGen v2 mistakenly gave the capybara human hands and placed the passenger in the front seat, overlapping with the steering wheel.
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 [dev]
- + Perfect text rendering with zero spelling errors.
- + Sophisticated gothic aesthetic with detailed thorns and cobwebs.
- + Professional layout that balances all prompt elements cleanly.
- − The parchment texture is concentrated mostly on the outer edges.
OmniGen v2
- + Vibrant color palette with high contrast.
- + Good use of backlight to create a silhouette effect for trees.
- − Significant text errors and garbled characters in the banner and details.
- − The border lacks the requested thorns, using a simple gold frame instead.
- − Messy composition at the bottom with overlapping and repetitive text.
Verdict: FLUX.2 [dev] followed every instruction perfectly, producing a polished invitation with high-fidelity text and authentic gothic details. OmniGen v2 struggled significantly with the text rendering and failed to include the thorns requested in the border, resulting in a cluttered and unusable design. FLUX.2 is the clear winner for its superior prompt adherence and typography.
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 [dev]
- + Excellent adherence to the isometric diorama perspective
- + Highly accurate text rendering and flag iconography
- + Nuanced, realistic PBR materials, especially for the rice and fish textures
- − Lighting is slightly muted compared to typical 'cartoon' aesthetics
OmniGen v2
- + Vibrant, appealing color palette
- + Good 3D block lettering for the 'JAPAN' text
- − The flag icon is incorrect
- − Sushi anatomy is nonsensical, featuring a tail-like structure on a roll
- − Text rendering is slightly inconsistent with the flag's position
Verdict: FLUX.2 [dev] followed all instructions perfectly, providing a clean isometric diorama with realistic textures and an accurate Japanese flag. OmniGen v2 failed on the details, producing an incorrect flag and unusual sushi structures that do not resemble real nigiri or maki.
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 [dev]
- + Excellent anatomical realism and fur texture rendering.
- + Includes all four requested animals correctly.
- + Beautifully soft, natural lighting with convincing god rays and dew effects.
- − Added an extra second rabbit not specifically requested.
OmniGen v2
- + Bright, vibrant colors that evoke a cheerful mood.
- + Clear focus on the center subjects.
- − Failed to include the bunny entirely.
- − Style is illustrative and cartoonish rather than 'hyper-photorealistic'.
- − Poor anatomy, particularly on the fox/kitten hybrid on the right.
Verdict: FLUX.2 [dev] successfully followed the complex prompt, delivering a highly detailed and realistic scene with all requested animals and atmospheric lighting. OmniGen v2 failed the realism requirement, producing a cartoonish image that missed the rabbit and featured inconsistent animal anatomy.
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 [dev]
- + Excellent adherence to the 'soft pastel' and 'hand-painted' style requirements.
- + Captures the specific Ghibli character aesthetic (e.g., The Wind Rises) with high fidelity.
- + Preserves the nuanced facial expressions, especially the girlfriend's look of disapproval.
- − Replaces the urban background with a generic flower field, altering the original context.
- − The red dress has become significantly more pink, slightly losing the original color identity.
OmniGen v2
- + Successfully preserves the urban environment and the red color of the dress.
- + Effective translation of the meme's composition into a clean anime style.
- + Good preservation of background details like the red bus and pedestrians.
- − The style feels more like modern generic anime than the specific 'hand-painted' Ghibli aesthetic requested.
- − Loss of the specific 'shocked' expression of the girlfriend, who now looks content.
- − Colors are bold and saturated rather than the requested soft pastels.
Verdict: FLUX.2 [dev] followed the stylistic instructions much more closely, delivering a beautiful hand-painted watercolor look that perfectly matches the Studio Ghibli prompt, though it changed the background. OmniGen v2 preserved the original image's composition and setting better but failed to capture the requested artistic style and the key emotional expressions of the subjects. FLUX.2 [dev] is the preferred choice for its superior artistic execution and mood.
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 [dev]
- + Successfully added a large volume of realistic leaves with varied motion.
- + Preserved the identity of the woman and the dog with high fidelity to the source image.
- + The motion blur on the leaves and strands of hair creates a strong sense of wind.
- − The hair effect is slightly symmetrical, giving it a somewhat artificial 'fan' look.
OmniGen v2
- + The hair motion looks more naturally one-directional, suggesting a specific wind current.
- + The overall image colors and lighting remain pleasant and consistent.
- − The leaves are very few and look like static clipart rather than dynamic objects.
- − Failed to preserve the source image details, significantly altering the woman's face and the dog's appearance.
- − The editing precision is low, resulting in a generic recreation rather than a modification of the original.
Verdict: FLUX.2 [dev] performed significantly better as an image editor by maintaining the original subjects' identities while effectively integrating dynamic motion through realistic leaves and wind-blown hair. OmniGen v2 failed to preserve the source image's specific details, effectively generating a new, similar image that lacked the 'energetic' feel requested due to its static-looking leaf additions.
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 [dev]
- + Excellent typography with perfect spelling and correct accent mark.
- + Professional vector emblem style with sophisticated line work.
- + Subtle grainy texture adds a high-quality vintage feel.
- − The Cloche is slightly less prominent than in the other model.
OmniGen v2
- + Strong minimalist silhouette that is very legible.
- + Includes the requested steam and banner elements clearly.
- − Major spelling error in the restaurant name.
- − Text spacing is uneven, and the banner ends don't align properly with the curve.
- − Lacks the requested subtle texture, appearing a bit flat.
Verdict: FLUX.2 [dev] successfully followed all prompt instructions, including complex text rendering with correct spelling and accents, while maintaining a professional vector aesthetic. OmniGen v2 failed on text accuracy ('CAFFFLORIN') and provided a much simpler composition that lacked the refined texture asked for in the prompt. FLUX.2 [dev] is the clear winner for its superior graphic design quality and technical precision.
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 [dev]
- + Strong prompt adherence with clearly defined icons for launch, orbit, and landing.
- + Modern flat-vector aesthetic with a clean navy and white color palette.
- + Includes relatively accurate depictions of the Saturn V and Lunar Module.
- − The layout order is jumbled, not following a linear 1-6 numerical sequence.
- − Contains several OCR spelling errors in the labels like 'Trarquility' and 'Ualsist Ordit'.
OmniGen v2
- + Features a very clean, structured grid layout characteristic of an infographic.
- + Adheres strictly to the requested NASA color palette.
- − Failed the primary prompt instruction by labeling it 'APOLO 17' instead of Apollo 11.
- − Icons are generic and repetitive, failing to represent the specific stages like the Saturn V or trajectory.
- − Severe text rendering issues and gibberish throughout the labels.
Verdict: FLUX.2 [dev] is the clear winner as it successfully illustrated the complex specific elements requested, such as the Saturn V and the Lunar Module stages, despite some jumbled ordering and minor typos. OmniGen v2 failed the most basic prompt requirement by targeting the wrong mission number (Apollo 17) and used extremely generic icons that do not convey the infographic's intended steps.
Explore each model
Unified multimodal model for text-to-image generation, instruction-guided image editing, personalized generation, and virtual try-on