HiDream I1 Full HiDream AI vs Qwen Image Alibaba

HiDream I1 Full

• + Excellent depiction of light and shadow, with convincing soft window lighting.
• + Beautiful texture on the wooden table and the red book cover.
• + Realistically distorted refraction of the plant through the glass cube.

• − The blue sphere appears to be floating rather than sitting on the bottom surface.
• − Some ghosting artifacts are present within the glass refraction.

Qwen Image

• + The blue sphere is firmly grounded within the cube.
• + Clean, minimalist composition with sharp edges on the glass.
• + Accurate placement of all elements requested in the prompt.

• − The lighting feels a bit more flat and artificial compared to Model A.
• − The plant visibility through the glass is slightly less realistic than in Model A.

HiDream I1 Full

• + Excellent shallow depth of field and bokeh
• + Subtle, realistic skin texture on the man's face
• + Beautifully rendered wet pavement with clear mirror-like reflections

• − Failed the motion blur request as the passing cars appear frozen
• − Anatomical issues with the man's sitting position and the bicycle's proportions

Qwen Image

• + Successfully incorporated motion blur on the background vehicle
• + Achieved a more realistic 'imperfect' snapshot framing
• + The act of repairing the bicycle looks more natural

• − Visible rain streaks look like a filter rather than a natural atmospheric effect
• − Missing the requested 'shallow depth of field' as the cars are sharp despite the motion blur
• − Generic skin texture compared to the detail in the other model

HiDream I1 Full

• + Excellent symmetrical framing and intense facial expression
• + Superior rendering of 'lifelike eyes' as requested
• + Detailed and believable metallic textures with realistic aging

• − The braids at the top appear somewhat flat/unnatural in their integration
• − Lighting is a bit diffused rather than strong warm torchlight

Qwen Image

• + Successfully included colorful beads in the hair braids
• + Strong use of warm torchlight and visible bokeh sparks
• + High quality texture on the leather straps and buckles

• − The sparks have a very artificial, 'stamp' like appearance
• − The facial scars look a bit like paint rather than biological tissue
• − The composition feels slightly cluttered on the right side

HiDream I1 Full

• + Excellent text legibility for main headings like 'RESTAURANT MENU' and 'MAINS'.
• + Clear, high-resolution food photography that looks appetizing and realistic.
• + Good use of bold sans-serif fonts which aligns well with the 'modern' part of the prompt.

• − The grid is a bit repetitive, using four very similar pizza images instead of a variety of foods.
• − The 'Appetizers / Pizza' section has no text-based menu items beneath it.

Qwen Image

• + Perfectly captures the 'colorful grid' and 'vibrant accents' request with multi-colored backgrounds for food photos.
• + More creative layout that feels like a complete professional menu design with a distinct sidebar for photos.
• + Better variety of food shown, including salads and slices, not just whole pizzas.

• − The text is largely illegible 'lorem ipsum' style gibberish compared to Model A.
• − The word 'Pizzaurant' is a bit of an odd creative choice for a general restaurant menu request.

HiDream I1 Full

• + Successfully rendered the correct date (April 30, 2026)
• + Excellent chalk texture and realistic wood grain on the frame
• + Correctly interpreted an elegant cursive style for the title

• − Completely failed to render the specific menu items requested, generating gibberish text instead
• − Text looks more like a digital font than natural handwriting with variations
• − Included an unnecessary illustration of a cup

Qwen Image

• + Successfully rendered the specific requested menu items with much higher accuracy
• + Captures a better 'handwritten' feel with variations in letter height and spacing
• + Composition fits the cozy café prompt well

• − Major error in the year, rendering '20026' instead of '2026'
• − Spelling error in 'Risoto' (missing one 's')
• − The chalk texture is slightly too smooth and looks a bit like a digital marker

HiDream I1 Full

• + Excellent sense of motion and dynamic composition.
• + Highly detailed textures on both the space suit and the horse's coat.
• + Strong cinematic lighting with vibrant reflections on the helmet.

• − The horse's front legs have slightly awkward anatomy at the joints.
• − Failed the specific negative constraint to have the 'horse on top' of the astronaut.

Qwen Image

• + Natural and balanced composition with a clear view of the planet.
• + Clean rendering of the astronaut's gear and harness details.
• + Good color contrast between the brown horse and the blue Earth.

• − Failed the specific negative constraint to have the 'horse on top' of the astronaut.
• − The perspective of the planet in the background feels slightly flat compared to the foreground.

HiDream I1 Full

• + Excellent text rendering on the hat.
• + Better composition with a more dynamic close-up angle.
• + The lighting and skin tones of the passenger look more realistic.

• − The paws on the steering wheel look somewhat distorted and less animal-like.
• − Perspective on the steering wheel pillar is slightly off.

Qwen Image

• + Successfully rendered paws that look like actual animal hands gripping the wheel.
• + Includes a taxi sign on the roof which adds to the requested realism.
• + The driver's cap style is more authentic to a formal chauffeur/taxi driver uniform.

• − The text on the roof sign is garbled ('YOXI').
• − The passenger's facial features and eyes look slightly less refined compared to Model A.
• − The steering wheel placement feels slightly disconnected from the driver.

HiDream I1 Full

• + Excellent graphic design layout with a clear focal point
• + Successful rendering of the scroll banner requested in the prompt
• + High clarity and contrast in the illustration elements

• − Fails significantly on text accuracy for the event details
• − Omits the word 'Invitation' from the main title
• − The location and specific date are completely hallucinated as gibberish

Qwen Image

• + Near-perfect adherence to specific event text (Date, Time, Location)
• + Great atmosphere with cinematic lighting and twisted trees
• + Includes both the thorn and spiderweb border elements as requested

• − Main title text 'Halloween' is slightly malformed
• − The scroll banner has a visual glitch where text is split awkwardly
• − The thorny border looks a bit repetitive and digital

HiDream I1 Full

• + Excellent text rendering with clean, bold typography.
• + High-quality textures on the fish and roe that give a premium 3D look.
• + Clearer resolution and sharper details on the sushi components.

• − Missed the small flag icon requested in the prompt.
• − The background contains some slight digital noise or texture instead of being a perfectly solid fill.

Qwen Image

• + Successfully included the small flag icon as requested.
• + Features more detailed props like chopsticks and a physical flag pole that enhance the diorama feel.
• + Better adherence to the 'small raised diorama' aspect of the prompt.

• − Lower image clarity with softer, almost blurry edges compared to Model A.
• − The rice grains look like large uniform spheres, which feels slightly less refined than Model A's texture.

HiDream I1 Full

• + Excellent fur texture on the fox and golden retriever.
• + Strong adherence to the 'god rays' and 'golden sunrise light' prompt.
• + Vibrant and saturated color palette that feels very joyful.

• − Completely missed the 'baby bunny' requirement, providing a second kitten instead.
• − The style is more illustrative and stylized than 'hyper-photorealistic'.
• − The puppy's eyes look a bit artificial and cartoonish.

Qwen Image

• + Successfully included all four requested animals (dog, cat, bunny, and fox).
• + Achieved a much more photorealistic look with believable eye reflections.
• + Better captures the active 'chasing' and 'tumbling' motion described in the prompt.

• − The kitten has five paws/limbs visible, representing a significant anatomical error.
• − The puppy's head is slightly large and floating-looking relative to its body.

HiDream I1 Full

• + Clean vector aesthetic that matches the minimalist prompt.
• + Text in the banner is perfectly legible and centered.
• + Accurate interpretation of the 'Est. 1720' requirement.

• − Completely missed the primary text 'Caffè Florian'.
• − The coffee cup inside the cloche causes a confusing visual overlap with the cloche's handle.

Qwen Image

• + Includes all requested text elements including 'Caffè Florian'.
• + Good use of subtle paper texture on the background.
• + More dynamic cloche design with steam coming out of the top.

• − The main text 'Florian' has severe kerning and overlapping issues, making it messy.
• − The banner proportions are a bit chunky compared to the elegant vintage prompt.

HiDream I1 Full

• + Strong aesthetic appeal with a professional vector illustrative style.
• + Excellent adherence to the NASA-inspired color palette.
• + Crisp text for the main header.

• − Fails to include all 6 requested steps, stopping after 3-4 loosely defined sections.
• − Text content for the sub-steps is nonsensical and includes prompt keywords like 'ICON'.
• − Included a space shuttle silhouette and a Saturn-like planet which are historically inaccurate for Apollo 11.

Qwen Image

• + Follows the logical flow of the requested 6 steps much more accurately.
• + Successfully visualizes the lunar module landing component requested.
• + Includes a reasonable representation of the mission crew with recognizable names.

• − The 'Descent' and 'Landing' icons are swapped or incorrectly numbered.
• − Visual style is slightly less 'clean' and lacks the balance of professional infographic design.
• − Minor spelling errors in labels like 'Tranar' and 'Aldin'.