FLUX.1 Kontext [max] Black Forest Labs vs HiDream I1 Full HiDream AI

FLUX.1 Kontext [max]

• + Excellent adherence to lighting instructions with realistic caustics and window shadows.
• + Very sharp textures on the blue sphere and wooden table.
• + Perfect logical placement of objects within the scene.

• − The plant is more overhead than strictly behind the cube.
• − The blue sphere has a slightly rough, crystalline texture rather than being a smooth ball.

HiDream I1 Full

• + Successfully positions the plant directly behind the glass cube.
• + Clean, minimalist composition with smooth object surfaces.

• − The blue sphere is levitating inside the cube, which feels physically unnatural.
• − The glass thickness and refractions are less realistic compared to Model A.
• − Lighting feels a bit more generic despite the visible window.

FLUX.1 Kontext [max]

• + Excellent depiction of falling rain and realistic wet textures.
• + The bicycle mechanics and the man's interaction with the chain are physically plausible.
• + Successfully captures the 'imperfect framing' and cinematic bokeh of a 50mm lens.

• − The man's heritage appears more ambiguous than specifically Japanese.
• − The scale of the bicycle seems slightly oversized compared to the crouching man.

HiDream I1 Full

• + Stronger adherence to the requested Japanese ethnicity for the subject.
• + Effective shallow depth of field and beautiful ground reflections.
• + Good composition with the city background providing depth.

• − The bicycle geometry is broken, with the frame passing through the man's leg.
• − The man is sitting on the frame or a phantom seat rather than 'repairing' it actively.
• − Does not show visible falling rain, only wet surfaces.

FLUX.1 Kontext [max]

• + Naturalistic lighting and highly realistic skin texture
• + Excellent engraving details on the plate armor
• + Subtle and organic integration of braids and hair fibers

• − The sparks look slightly like flat streaks rather than glowing embers
• − Misses the specific request for beads in the hair

HiDream I1 Full

• + Strong adherence to the 'beads in hair' and 'faint scars' prompt elements
• + Good composition with a clear torch visible in the background
• + Distinct texture on the leather straps and metal buckles

• − Lighting feels overly artificial and high-contrast
• − Armor engraving looks a bit generic and repetitive compared to Model A
• − Facial scars look like surface-level paint rather than integrated skin texture

FLUX.1 Kontext [max]

• + Excellent professional layout with a clean graphic design feel
• + Strong grid-based composition that makes the menu feel organized
• + Includes a logo and diverse vibrant accents like the yellow footer

• − The text is mostly gibberish with inconsistent fonts for section headers
• − Food variety is lacking as most photos are just small variations of the same pizza

HiDream I1 Full

• + Successfully included specific headers for 'Appetizers / Pizza' and 'Mains'
• + Higher legibility of central text elements
• + Better color saturation in the food photos

• − Very basic layout that lacks the 'professional' feel of a modern menu
• − Alignment issues with the text overlapping images slightly at the bottom
• − Repetitive food photos despite having different category labels

FLUX.1 Kontext [max]

• + Perfect text accuracy for all requested menu items and prices.
• + Highly realistic chalk texture with smudge marks and natural handwriting styles.
• + Accurately interpreted the cutoff prompt to complete 'Brown Butter Chocolate Chip Cookies'.

• − The title uses a print-style block letter rather than the requested elegant cursive.
• − The background cafe details are slightly blurry and less defined than the board itself.

HiDream I1 Full

• + Elegant cursive title at the top matches the stylistic request better than Model A.
• + High visual clarity and clean composition with nice chalk illustrations.

• − Complete failure to follow the specific menu item text requested, resulting in gibberish.
• − The text appears too much like a digital font rather than natural handwriting.
• − Failed to include the specific pricing from the prompt.

FLUX.1 Kontext [max]

• + Excellent anatomical detail and texture on the horse.
• + Consistent lighting between the astronaut and the environment.
• + Clean, high-resolution rendering without significant artifacts.

• − Failed the negative constraint; the astronaut is riding the horse instead of vice versa.
• − The composition is a bit centered and standard for this prompt.

HiDream I1 Full

• + Dynamic composition with the background planet and clouds.
• + Good use of color contrast between the white horse and the dark space.

• − Failed the negative constraint regarding the position of the character and animal.
• − The back legs of the horse are anatomically distorted.
• − Low-quality texture on the planet surface and clouds.

FLUX.1 Kontext [max]

• + Excellent photorealistic fur texture and lighting
• + Cinematic composition with realistic background bokeh
• + Accurate depiction of the passenger as requested

• − The capybara only has one visible paw on the wheel instead of both
• − Composition feels slightly cramped with the side of the car blocking the view

HiDream I1 Full

• + Successfully placed both paws on the steering wheel as requested
• + Clearer view of the entire scene including the passenger and phone
• + Well-rendered 'TAXI' text on the cap

• − The passenger's proportions feel slightly off compared to the seat
• − Lighting is a bit more artificial and 'HDR' style compared to image A
• − Anatomical weirdness where the capybara's right arm seems to merge with its body

FLUX.1 Kontext [max]

• + Excellent text rendering with no spelling errors.
• + Perfect adherence to all specific event details included in the prompt.
• + Atmospheric and polished cinematic lighting that fits the 'gothic' theme.

• − The 'Arches, NYC' location is repeated twice at the bottom.
• − The background is very dark, making the twisted trees slightly difficult to see.

HiDream I1 Full

• + Strong composition with a clear parchment paper effect in the center.
• + Good use of the spiderweb and thorn border elements described.

• − Failed significantly on text accuracy, producing gibberish like 'Yootmber' and '30 cm'.
• − Missing the word 'Invitation' from the main title.
• − The overall style is more 'cartoony' than the requested 'cinematic' and 'vintage gothic' look.

FLUX.1 Kontext [max]

• + Successfully added a full head of hair that looks natural and thick.
• + Excellent preservation of the original facial identity and expression.
• + Maintains the lighting and background of the source image perfectly.

HiDream I1 Full

• + Successfully changed the hairstyle to a mullet-style length at the back.

• − Completely changed the background and environment.
• − Introduced severe artifacts like a bizarre black object protruding from the beard.
• − Modified the facial features, making the person unrecognizable from the original.

FLUX.1 Kontext [max]

• + Excellent 3D soft textured aesthetic that perfectly matches the 'cartoon scene' prompt.
• + Accurate text rendering with a stylized font that fits the theme.
• + High-quality material rendering on the sushi and diorama base.

• − Missing the requested flag icon.
• − The 'diorama base' is a bit plain compared to Model B.

HiDream I1 Full

• + Stronger isometric diorama composition with more variety in sushi pieces.
• + Very clean, bold white typography.
• + Good adherence to the 45-degree top-down perspective.

• − Missing the requested flag icon.
• − The text 'SUSHI' is relatively small and less impactful than Model A.
• − Lighting feels slightly more flat and less 'refined' than Model A.

FLUX.1 Kontext [max]

• + Excellent adherence to all prompts, including a clear news desk, a microphone, a dog, and a hockey stick.
• + Captures the 'caricature' art style perfectly with exaggerated features and vibrant colors.
• + Preserves the subject's outfit (denim shirt and black top) accurately from the source.

• − The hockey stick is positioned awkwardly in the top right corner without much context.

HiDream I1 Full

• + Successfully creates a stylized caricature version of the person in the source image.
• + Includes a dog and preserves the facial structure of the original subject well.

• − Fails to include any clear elements of a TV show anchor (no microphone, desk, or studio).
• − Completely misses the 'hockey' requirement of the prompt.
• − The 'hockey puck' on the dog's mouth looks more like a weird mouthpiece or gag.

FLUX.1 Kontext [max]

• + Successfully included all four requested animals (dog, cat, rabbit, fox).
• + Excellent rendering of soft, backlit fur and natural meadows textures.
• + Consistent lighting and coherent 'god rays' following the sun's position.

• − The rabbit's hands/paws are slightly anatomically awkward.

HiDream I1 Full

• + Vibrant colors and a very cute, symmetrical composition.
• + High contrast and sharp focus on the central animals.

• − Missed the rabbit entirely, instead providing two kittens.
• − Animals look more like digital illustrations than the requested 'hyper-photorealistic' style.
• − The sun/glare is overly blown out and lacks detail.

FLUX.1 Kontext [max]

• + Excellent preservation of the original subjects' poses, clothing patterns, and composition.
• + Perfectly captures the Studio Ghibli aesthetic with soft watercolor textures and gentle line work.
• + Maintains the specific facial expressions of the 'distracted boyfriend' meme characters while stylizing them.

• − The background is slightly more washed out than typically seen in lush Ghibli films.

HiDream I1 Full

• + Successfully adopts a cute anime art style with vibrant colors.
• + Adds creative details like floral patterns and a hat that fit a summer anime theme.

• − Fails to preserve the core composition and distinct character dynamics of the source image.
• − The man's beard and the woman's hair style are significant departures from the original photographic subjects.
• − Loss of the iconic 'shocked' expression on the girlfriend character.

FLUX.1 Kontext [max]

• + Excellent preservation of the source subject and background.
• + Accurately adds motion to hair while maintaining the woman's identity.
• + Includes the dog and leash from the original image.

• − The added leaves are small and relatively sparse.
• − The left arm positioning is slightly altered, becoming stiff.

HiDream I1 Full

• + Great sense of dynamic motion through hair and clothing pose.
• + Highly energetic and lively interpretation of the prompt.

• − Fails to preserve the source image, completely removing the dog.
• − Changes the woman's facial features and identity significantly.
• − Background details like the flowers and bridge have been lost or changed.

FLUX.1 Kontext [max]

• + Excellent typography including the circumflex accent on 'Caffè'
• + Superior texture that gives a genuine vintage paper feel
• + Complete adherence to all prompt elements including the specific name.

• − The steam is a bit small relative to the cloche size.

HiDream I1 Full

• + Clean vector-style execution
• + Good use of the 'Est. 1720' banner

• − Completely failed to include the brand name 'Caffè Florian'
• − The steam appearing through the solid cloche creates a confusing visual logic
• − Lacks the requested 'subtle texture' compared to Model A.

FLUX.1 Kontext [max]

• + Excellent typography rendering for 'APOLLO 11' and the astronaut names.
• + Matches the requested navy, white, and muted red NASA palette perfectly.
• + Includes clever vector silhouettes of the three astronauts at the bottom.

• − Failed to provide all 6 specific steps, missing 'Launch' and 'Descent'.
• − The logical flow of the infographic is confusing, with arrows pointing in multiple directions.
• − The rocket icon looks like a generic toy rather than a Saturn V.

HiDream I1 Full

• + Successfully captured more of the requested infographic steps with clear labels.
• + The Saturn V rocket icon is much more accurate to the actual mission hardware.
• + Strong use of high-contrast flat vector style that feels like professional graphic design.

• − Contains significant text errors like '+ UNG' and '+ ICON' which appear to be hallucinations of the prompt text.
• − Composition is a bit cluttered with overlapping elements in the center.
• − Failed to include the lunar module landing and descent phases.