FLUX.1 [schnell] FP8 Black Forest Labs vs GPT Image 1 OpenAI

FLUX.1 [schnell] FP8

• + Excellent handling of light refraction and reflections on the wood surface.
• + Realistic blue globe with intricate glass textures.
• + Follows the lighting direction prompt perfectly with bright window light.

• − The 'cube' is a tall rectangular prism rather than a cube.
• − The internal shelf is an hallucinated element not mentioned in the prompt.

GPT Image 1

• + Accurately depicts a cubic shape for the glass container.
• + Highly realistic texture for the matte blue sphere and the paper edges of the book.
• + Shows the plant clearly behind the glass as requested.

• − The sphere appears to be floating inside the cube without touching the bottom surface.
• − The lighting is a bit more muted than the 'window light' description suggests compared to Image A.

FLUX.1 [schnell] FP8

• + Excellent handling of wet pavement reflections and light artifacts
• + Detailed mechanical rendering of the bicycle frame
• + Good sense of environmental depth in an urban setting

• − The man appears to be posing with or leaning on the bike rather than 'repairing' it
• − Fails to significantly incorporate the requested motion blur for passing cars
• − The subject's ethnicity is somewhat ambiguous

GPT Image 1

• + Strong adherence to the 'repairing' action with a realistic crouching pose
• + Excellent skin texture and authentic elderly Japanese features
• + Successfully captures the moody, desaturated color palette and cinematic atmosphere

• − The bicycle geometry becomes physically impossible/muddled around the rear wheel hub
• − Missing clear 'motion blur' on the cars in the background as requested

FLUX.1 [schnell] FP8

• + Extremely sharp and detailed skin textures
• + Intense and dramatic lighting
• + Lifelike iris detail and coloring

• − Hair lacks the requested braided style with beads
• − The facial expression feels slightly over-dramatized/caricatured
• − Subtle text artifacts visible on the bottom right armor section

GPT Image 1

• + Accurately depicts hair braided with small beads
• + Ornate engraved plate armor is highly detailed and realistic
• + Captures the battle-worn aesthetic with believable dirt and scars

• − Lighting is a bit darker, obscuring some details
• − The skin texture is slightly softer compared to the other image

FLUX.1 [schnell] FP8

• + Comprehensive multi-section layout with many items
• + Great use of a grid for multiple food photos
• + Captures the professional, spread-out feel of a real physical menu

• − Numerous spelling errors in headings like 'APPTIZERS' and 'ORCETERS'
• − Food photography is small and slightly repetitive in style

GPT Image 1

• + High-quality food photography with vibrant colors
• + Clean, bold sans-serif typography that is very legible
• + Modern minimalist aesthetic with effective colored accents

• − Failed to include all requested sections (missing Mains)
• − Contains placeholder text like 'Apperoiation descrigion'

FLUX.1 [schnell] FP8

• + Excellent photorealistic rendering of the burger textures.
• + Good use of actual fire in the background for atmospheric depth.
• + Dynamic 'exploded' feeling with small debris and sauce droplets.

• − Significant text errors including 'LIIMITED' and 'NEEY'.
• − Incorrect price of €69 displayed prominently.
• − The primary burger is mostly assembled rather than fully exploded components.

GPT Image 1

• + Perfect text rendering for all requested phrases including the fiery effect.
• + Full adherence to the 'exploded' component layout showing individual layers.
• + Highly cohesive color palette with consistent glowing embers.

• − The price is rendered as '€.99' instead of '€6.99'.
• − The bun texture is slightly less realistic than Model A.
• − Lighting on the burger is somewhat flat compared to the intensity of the background.

FLUX.1 [schnell] FP8

• + Excellent chalk-like texture with smudges
• + Realistic wood frame and cafe environment context
• + Captures the casual, messy nature of handwriting

• − Terrible text accuracy with many misspellings like 'Risortto' and 'Octemon'
• − Fails to render the elegant cursive requested for the title
• − Repeats menu items multiple times in a confusing layout

GPT Image 1

• + Excellent text accuracy with no spelling errors
• + Consistent and clear chalk texture across all letters
• + Followed the specific menu item list and pricing accurately

• − Failed to provide 'elegant cursive' for the title
• − The handwriting looks a bit too uniform/digital rather than having 'natural variations'
• − The layout is very centered and static compared to a natural chalkboard

FLUX.1 [schnell] FP8

• + Successfully interpreted the surreal instruction of placing the horse on top of the astronaut equipment.
• + Cinematic lighting with a beautiful Earth backdrop.
• + High resolution and clear textures on the horse's fur and the mechanical gear.

• − Anatomy is a bit confused with two horse heads/bodies merged into the scene.
• − The 'astronaut' is represented as a piece of equipment rather than a person in a suit.

GPT Image 1

• + Excellent lighting and texture on the horse and space suit.
• + Great composition that feels very cinematic and illustrative.
• + Clear and coherent rendering of a horse and astronaut.

• − Failed the core prompt instruction to have the 'horse on top' of the astronaut.
• − Visual interpretation is literal and lacks the requested surrealism.

FLUX.1 [schnell] FP8

• + Excellent clarity and lighting contrast.
• + Accurately depicts the taxi interior and the 'bored' expression of the passenger.
• + High detail in the capybara's fur and clothing.

• − The passenger is holding two phones, which is a logic error.
• − The capybara's hat is a simple plastic-looking helmet rather than a traditional driver cap.

GPT Image 1

• + Features a more authentic taxi driver cap with a brim.
• + The passenger's bored expression perfectly matches the prompt's intent for normalcy.
• + Better depth of field and more realistic night street bokeh through the rear window.

• − The passenger is slightly out of focus compared to Image A.
• − The steering wheel placement is a bit low and crowded near the capybara's chest.

FLUX.1 [schnell] FP8

• + Features a vibrant, glowing jack-o-lantern central to the composition.
• + Includes stylized border flourishes that frame the image well.

• − Numerous spelling and grammar errors in the text like 'Tine', 'nigh tof friights', and 'Theaches'.
• − The text layout is messy with random colons and fragmented sentences.
• − Fails to correctly list the requested date and location clearly.

GPT Image 1

• + Excellent typography with a perfect 'Halloween Party Invitation' title and banner text.
• + Superior adherence to the vintage gothic aesthetic with a dark, moody parchment texture.
• + High-quality details such as the spider webs in the corners and the subtle moon and bat silhouettes.

• − Combined the Time and Location fields into one line ('TIME: The Arches, NYC'), missing the '7pm' detail.

FLUX.1 [schnell] FP8

• + Excellent 3D rendering with soft shadows and realistic textures
• + Accurate miniature diorama aesthetic
• + High clarity and clean composition

• − Text is incorrect and fails to generate 'SUSHI'
• − Includes a weird second 'JAPAN' word with a dot instead of the requested layout

GPT Image 1

• + Perfect text adherence with both 'JAPAN' and 'SUSHI' and a flag icon
• + Accurate isometric diorama perspective and lighting
• + Clean, high-quality cartoon style with pleasant textures

• − Background color is slightly darker than 'light blue' requested
• − The rice texture is a bit large for the scale

FLUX.1 [schnell] FP8

• + Includes a high number of animals and butterflies
• + Vibrant golden lighting and colors

• − Failed to include a rabbit, replacing it with extra kittens/unknown kits
• − Anatomical glitches including a floating paw and a fifth animal head emerging from the fox
• − Looks more like a digital illustration than a photorealistic scene

GPT Image 1

• + Perfect adherence to the prompt, including all four specific animals requested
• + Excellent movement and dynamic composition that captures the 'chasing' aspect
• + Realistic lighting, fur textures, and believable depth of field

• − One butterfly wing is slightly disconnected from its body
• − Fox paws are a bit dark and muddy in texture

FLUX.1 [schnell] FP8

• + Successfully captures the requested warm brown and cream tones
• + Includes a decorative banner and light background as requested

• − Significant spelling error in the primary brand name
• − The dome looks more like a building cupola than a food cloche
• − Composition is cluttered with overlapping text

GPT Image 1

• + Perfect text rendering of the name and date
• + Accurate depiction of a minimalist retro cloche dome with steam
• + Excellent minimalist aesthetic and clean vector style

• − Ignored the 'light background' instruction, opting for black
• − Used a single brown tone instead of a brown and cream color scheme

FLUX.1 [schnell] FP8

• + Strong composition that feels like a professional corporate infographic
• + Excellent use of a dark navy palette with clean, sharp vector lines
• + Attempts to include almost all requested complex text and numbered steps

• − Text is mostly gibberish with frequent spelling errors
• − Icons are abstract and don't clearly represent the Saturn V or specific mission stages

GPT Image 1

• + Text is highly legible and correctly identifies the crew members
• + Icons for Earth and the Lunar Module are very clear and recognizable
• + Excellent flat vector illustration style with a cohesive color palette

• − Fails to follow the requested 6-step linear structure, combining/missing some phases
• − Layout feels a bit cluttered compared to a traditional infographic timeline
• − Contains some typos like 'EARLLUNAR'