Black Forest Labs' open-weights image generation model with frontier performance, available for non-commercial local deployment
Settled by community votes across 16 shared challenges, with an AI judge weighing in on each.
FLUX.2 [dev]
#19 of 62 in Text-to-Image
HiDream I1 Full
#60 of 62 in Text-to-Image
Where the votes landed
FLUX.2 [dev]
0%
win rate
Ties
0%
HiDream I1 Full
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 adherence to lighting instructions with realistic soft window light.
- + Highly realistic textures on the wood table and the glass cube edges.
- + Perfect spatial arrangement of all requested objects.
- − The blue sphere's reflection on the bottom of the cube is slightly detached.
HiDream I1 Full
- + Successfully includes all prompt elements in the scene.
- + Good color vibrance on the blue sphere and red book.
- − The blue sphere is floating unnaturally in the center of the cube.
- − The perspective of the plant visible through the glass is slightly distorted.
- − Lower overall sharpness and texture detail compared to the other model.
Verdict: FLUX.2 [dev] produces a significantly more realistic and physically coherent image, with excellent textures and lighting that feels natural to the environment. HiDream I1 Full meets the prompt requirements but suffers from a floating sphere and less defined textures, making it look more like a digital render than a photograph.
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 skin and hair texture that looks highly realistic.
- + Successfully captures the motion blur from a passing car as requested.
- + The 'imperfect framing' feels like a genuine candid street photo.
- − The hands have some classic AI structural issues, particularly with finger merging.
- − The bicycle cables and mechanical parts are a bit tangled and nonsensical.
HiDream I1 Full
- + Beautiful reflections on the wet pavement.
- + Good color palette and atmospheric depth.
- − The man appears to be sitting on a miniature or structurally impossible bicycle.
- − Lacks the requested motion blur from passing cars.
- − The hands are severely distorted, and the physics of the character interacting with the bike is poor.
Verdict: FLUX.2 [dev] followed the prompt much more accurately, specifically capturing the 'motion blur from passing cars' and the 'imperfect framing' that gives it a realistic candid feel. HiDream I1 Full failed on the motion blur requirement and produced a bicycle that was scaled incorrectly relative to the man, making the scene look staged and unnatural.
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 photorealism in the skin texture and scars.
- + Ornate engraving on the plate armor is beautifully detailed and realistic.
- + Superior lighting integration with warm torchlight naturally interacting with all surfaces.
- − The beads in the hair are a bit small and less distinct than requested.
- − The leather straps lack significant wear compared to the facial scarring.
HiDream I1 Full
- + Very distinct and stylized hair beads that align well with the prompt.
- + The 'battle-worn' aesthetic is conveyed through grit and blood on the armor.
- − Over-sharpened visual quality leads to a 'digital' look rather than a lifelike one.
- − The eyes look slightly unnatural and overly saturated.
- − Anatomy of the neck and how the braids fall over the armor feels a bit forced.
Verdict: FLUX.2 [dev] produces a significantly more lifelike and cohesive image, with masterful lighting and realistic skin textures that match the 'paladin' archetype perfectly. HiDream I1 Full captures the specific details like hair beads more boldly, but the overall image suffers from over-processing and a less natural aesthetic compared to FLUX.2 [dev].
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 professional layout with distinct columns and headers.
- + High-quality, varied food photography that fits a restaurant aesthetic.
- + Clean use of vibrant accent colors as requested.
- − Text consists largely of illegible gibberish characters.
- − Incorrectly pluralized 'PIZZAU' and 'DIZZA' in headers.
HiDream I1 Full
- + Features legible (though misspelled) sans-serif English text.
- + Clean, minimalist white background with a simple grid.
- + Good balance of text and image spacing.
- − Limited variety in food photos, showing only pizzas for all sections including 'Mains'.
- − Text contains several typos like 'Chapne' and 'Surmper'.
- − Layout is quite basic and lacks the 'professional' polish of a real menu.
Verdict: FLUX.2 [dev] produces a much more realistic and professionally designed layout that captures the essence of a modern menu, despite the text being mostly gibberish. HiDream I1 Full has more legible text, but fails significantly on content variety by only showing pizzas even in the 'Mains' category, and the design feels amateurish by comparison.
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 prompt adherence with all text correctly spelled.
- + Highly realistic chalk texture with smudges and natural handwriting variations.
- + Authentic blackboard appearance with chalk dust and appropriate letter slant.
- − The layout is a bit cramped at the bottom, though legible.
- − Slight distortion in the word 'with' on the second item.
HiDream I1 Full
- + Clean, aesthetically pleasing layout and wood frame.
- + The title and date are rendered correctly and clearly.
- − Complete failure to follow prompt instructions for the menu items, producing gibberish text.
- − Text appears too smooth and digital, lacking authentic chalk texture.
- − All prices were hallucinated and do not match the prompt.
Verdict: FLUX.2 [dev] followed the prompt almost perfectly, correctly spelling the complex menu items and creating a convincing chalk-on-blackboard texture. In contrast, HiDream I1 Full failed significantly on prompt adherence, replacing the requested menu items with illegible gibberish and ignoring the specified prices.
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 photographic realism with detailed textures on the spacesuit and horse fur.
- + Sophisticated lighting and cinematic depth of field.
- + Natural interaction between the rider's hands and the reins.
- − Failed the specific spatial instruction; the astronaut is on top, not the horse.
HiDream I1 Full
- + Dynamic composition with a strong sense of movement.
- + Clear, high-contrast imagery that fits the 'cinematic' prompt.
- − Failed the specific spatial instruction; the astronaut is on top, not the horse.
- − The horse's front-right leg has an anatomical glitch where the hoof joins.
Verdict: Both models failed the 'trick' prompt instruction to place the horse on top of the astronaut, instead providing the standard astronaut-on-horse configuration. FLUX.2 [dev] is the superior image due to its significantly higher level of detail, realistic lighting, and better anatomical coherence compared to HiDream I1 Full.
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 'professional' expression for the capybara.
- + Logical camera placement from outside the windshield looking in.
- + Highly realistic textures on the capybara's fur and the passenger's coat.
- − The capybara's front paws appear more like human-monkey hybrid hands than natural capybara paws.
- − Minor clipping where the steering wheel and paws meet.
HiDream I1 Full
- + Clean and vibrant colors with clear 'TAXI' text on the hat.
- + Successfully captures the bored expression of the passenger in the back.
- + Good depth of field with the blurred Manhattan lights.
- − The perspective is awkward, with the passenger appearing too large and close to the driver for a standard taxi.
- − The capybara's head is fused oddly with the jacket collar.
- − Anatomy of the paws on the steering wheel is poorly defined.
Verdict: FLUX.2 [dev] is the superior model because it better captures the 'professional' demeanor requested and maintains a more realistic spatial relationship between the driver and the passenger. HiDream I1 Full has better text rendering on the hat but suffers from perspective issues that make the passenger appear to be sitting right next to the driver's seat.
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 for all requested details and dates.
- + Superior composition with a cohesive thorny border and moody atmospheric lighting.
- + Excellent parchment texture and thematic consistency.
- − The thorns are slightly busy at the top of the frame.
HiDream I1 Full
- + Good inclusion of bats and twisted trees in the background.
- + Strong contrast in the jack-o-lantern's glow.
- − Severely failed to render the specific date and location information correctly.
- − Text contains gibberish such as 'Yootmber' and '30 cm'.
- − Missing the word 'Invitation' from the main title.
Verdict: FLUX.2 [dev] followed every instruction perfectly, including complex text rendering for the date and location, while maintaining a high-quality vintage aesthetic. HiDream I1 Full failed significantly on the text requirements, producing nonsensical words and incorrect information despite a decent artistic style.
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 [dev]
- + Excellent preservation of the original face, clothing, and background.
- + Realistic hair texture and lighting integration.
- + Follows the instruction to add a full head of hair perfectly.
- − The hair style choice (large afro) is a bit dramatic compared to the original beard texture.
HiDream I1 Full
- + Added some hair to the back of the head.
- − Fails to preserve the original face, changing the subject entirely.
- − Completely changes the background environment from a desert to a field with windmills.
- − Introduces major artifacts like the strange black object across the face.
- − Does not provide a 'full head of hair' as the top remains bald.
Verdict: FLUX.2 [dev] successfully performed the edit by adding hair while keeping all other aspects of the source image identical. HiDream I1 Full failed on almost every metric, regenerating the entire image with a different person and environment while adding nonsensical artifacts.
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 flag icon requirement
- + Beautifully rendered PBR textures on the fish and wasabi
- + Clean, centered typography with high readability
- − The sushi and plate are slightly off-center compared to the diorama base
HiDream I1 Full
- + Strong 3D isometric perspective and composition
- + Vibrant colors that pop against the blue background
- + Accurate execution of the raised diorama base
- − Failed to include the requested Japanese flag icon
- − The 'SUSHI' text is slightly small and less prioritized
- − The rice grains look a bit like individual plastic beads rather than cohesive sushi rice
Verdict: FLUX.2 [dev] followed all the prompt instructions, including the specific requirement for a small flag icon, and provided superior texture quality that feels more realistic yet stylized. HiDream I1 Full produced a good isometric composition but missed the flag requirement and had slightly less refined text rendering.
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 [dev]
- + Successfully incorporates all elements including hockey, dogs, and the TV anchor desk.
- + Maintains a strong resemblance to the source subject while adopting a caricature style.
- + Extremely creative composition with dogs in jerseys and a hockey rink set.
- − Text on the desk is garbled and unreadable.
- − Includes a random secondary human figure that wasn't requested.
HiDream I1 Full
- + Captures the caricature facial style well and maintains the source subject's identity.
- + High visual clarity in the rendering of the character and the dog.
- − Completely ignores the 'hockey' part of the prompt.
- − The 'TV show anchor' element is very subtle, relying only on a blurred background TV rather than a clear profession depiction.
- − Anatomical issues with the dog merging into the shoulder.
Verdict: FLUX.2 [dev] is the clear winner as it successfully integrated every specific element of the prompt—TV anchor desk, hockey rink, and dogs—into a cohesive and humorous caricature. While HiDream I1 Full produced a clean caricature, it failed to include the hockey theme and had a much simpler composition that did not fully reflect the requested profession.
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 accuracy for all animal types.
- + Masterful handling of lighting, including god rays and dew sparkles in the grass.
- + Photorealistic fur texture and high environmental detail.
- − Includes two bunnies instead of one, slightly missing the specific count requested.
HiDream I1 Full
- + Captures a very 'wholesome' and cute aesthetic with expressive eyes.
- + Vibrant colors that pop.
- − Missing the baby bunny entirely.
- − Looks more like a digital painting or illustration than a 'hyper-photorealistic' scene.
- − The puppy's paws and the fox's body have blurred, incoherent blending with the grass.
Verdict: FLUX.2 [dev] significantly outperforms HiDream I1 Full by delivering a truly photorealistic image with sophisticated lighting and distinct textures, whereas HiDream I1 Full has a stylized, airbrushed look. Furthermore, FLUX.2 [dev] successfully included all species mentioned (despite adding an extra bunny), بينما HiDream I1 Full failed to include the bunny at all.
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 preservation of the source image's composition and character poses.
- + Perfectly captures the soft, pastel color palette requested in the prompt.
- + Highly accurate rendering of the iconic meme face expressions in an illustrative style.
- − The transition from a city street to a flower field is a significant background change, though it fits the 'dreamy' request.
HiDream I1 Full
- + Strong Studio Ghibli character design aesthetic with rounder features and blush.
- + Maintains the urban street setting of the original image.
- + Creative addition of patterns to the clothing that feel thematic.
- − Fails to preserve the core 'distracted boyfriend' pose, as the man is looking forward and smiling rather than looking back in surprise.
- − Loses the specific facial expressions (the girlfriend's anger, the man's longing) that define the source image.
Verdict: FLUX.2 [dev] is the clear winner as it successfully transforms the image into a beautiful Ghibli-inspired watercolor style while strictly preserving the composition and emotional context of the original meme. HiDream I1 Full adopts a recognizable anime style but fails as an edit because it alters the characters' poses and expressions, losing the 'distracted' narrative of the source image.
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]
- + Excellent source preservation, maintaining the woman's face, clothing, and the dog perfectly.
- + Strong prompt adherence, adding wind-blown hair and flying leaves.
- + Maintains consistent lighting and background details from the original image.
- − The wind effect on the hair is slightly symmetrical and looks a bit digital.
HiDream I1 Full
- + Successfully captures a high-energy, lively feeling as requested in the prompt.
- + Leaves and hair motion feel very integrated with the character's movement.
- − Very poor source preservation; the person has been changed, the dog is missing, and the background is altered.
- − The woman's right hand and feet have significant anatomical issues/distortions.
Verdict: FLUX.2 [dev] is the clear winner as it successfully follows the editing instructions while perfectly preserving the identity and core elements of the source image. In contrast, HiDream I1 Full essentially generated a new image, failing to preserve the specific woman or the dog from the original while introducing significant anatomical artifacts.
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]
- + Perfect text rendering including the accented 'è' in Caffè Florian.
- + Excellent banner work with complex, clean vector-style lines.
- + Authentic subtle texture on the background that mimics vintage paper.
- − The steam lines on top of the cloche are very small and somewhat disconnected from the main silhouette.
HiDream I1 Full
- + Creative interpretation of the cloche as a glass dome revealing a coffee cup.
- + Bold, clear visual hierarchy that is easy to read at a glance.
- − Completely failed to include the brand name 'Caffè Florian' which was the primary subject.
- − The cloche geometry is slightly asymmetrical on the left side curve.
- − Lacks the requested 'subtle texture' on the background.
Verdict: FLUX.2 [dev] followed every instruction perfectly, producing a professional-grade vector logo with precise typography and a vintage feel. HiDream I1 Full failed to include the actual name of the restaurant, making it unusable for the specific prompt, despite having an interesting visual concept.
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]
- + Successfully followed the palette and requested steps accurately
- + Higher quality icons with sophisticated vector shading
- + Excellent text readability for major headers
- − Included random garbled text at the bottom and in the center
- − Repeated some icons out of sequence
HiDream I1 Full
- + Strong composition with a clear central focal point
- + Good use of the navy, white, and red color palette
- − Failed to include 5 out of the 6 requested mission steps
- − Included literal words from the prompt like 'ICON' and '+ UNG' in the design
- − Inaccurate planetary depictions (Saturn-like rings on a red planet)
Verdict: FLUX.2 [dev] followed the complex instructions much better, providing a grid of icons that represented almost all the requested mission phases, despite some hallucinatory text. HiDream I1 Full missed the majority of the infographic steps and included prompt meta-text in the final image. FLUX.2 [dev] is the clear winner for its superior prompt adherence and sophisticated vector art style.
Explore each model
HiDream AI's 17B parameter text-to-image model using sparse diffusion transformer with mixture of experts, achieving state-of-the-art image generation quality with strong prompt following