FLUX.1 Kontext [dev] Black Forest Labs vs GPT Image 1 OpenAI

FLUX.1 Kontext [dev]

• + Excellent photographic realism and lighting
• + Accurate reflections and refractions through the glass
• + Correct placement of the plant behind the glass

• − The glass cube has an unusually thick mirrored or solid base

GPT Image 1

• + Perfect adherence to all spatial instructions
• + Clean and simple composition
• + Consistent glass texture throughout the cube edges

• − The sphere appears slightly matte, making it look a bit flat compared to the lighting
• − The plant is visible behind the cube but doesn't interact with the glass refraction as realistically as Model A

FLUX.1 Kontext [dev]

• + Excellent color vibrance and clear rainy-day atmosphere
• + Accurate representation of an elderly Japanese man

• − Man is posing with the bike rather than repairing it
• − The composition feels centered and staged, failing the 'imperfect framing' prompt
• − Cars in the background are stationary and lack motion blur

GPT Image 1

• + Perfectly captures the 'repairing' action requested
• + Authentic 'candid' feel with imperfect framing as requested
• + Detailed skin texture and realistic, muted color grading

• − The red of the bicycle is slightly less prominent than in the other image
• − The background motion blur is subtle rather than pronounced

FLUX.1 Kontext [dev]

• + Strong cinematic lighting with high contrast
• + Highly detailed and clearly engraved armor design
• + Excellent skin texture and sharp eye details

• − Hair is primarily combed back rather than braided as requested
• − Lacks obvious beads in the hair
• − The 'battle-worn' effect feels a bit clean and stylized

GPT Image 1

• + Excellent adherence to the 'braided with beads' requirement
• + Stronger 'battle-worn' aesthetic with realistic dirt and grittiness
• + Very fine engraving on the metal armor textures

• − Lighting is a bit flatter compared to the cinematic quality of the first image
• − The bokeh sparks are less defined and more streaky

FLUX.1 Kontext [dev]

• + Excellent grid layout with varied crops for a stylized Look.
• + High-quality, vibrant food photography with complex textures.
• + Strong bold typography that fits a modern aesthetic.

• − Text consists of heavy gibberish that lacks any legible English words.
• − The sections do not clearly separate into the requested appetizers/pizza/mains categories.

GPT Image 1

• + Clear and logical hierarchy for a restaurant menu with clear pricing and item names.
• + High adherence to section requirements (Appetizers, Pizza) and bold sans-serif fonts.
• + Minimalist aesthetic with vibrant color-coded accents.

• − Text contains minor spelling errors ('descrigion', 'Apperoiation').
• − Composition is very basic compared to more modern graphic design trends.

FLUX.1 Kontext [dev]

• + Strong, vibrant color palette that matches an advertising aesthetic.
• + Clear, bold typography with a professional layout.

• − Failed the core prompt instruction for an 'exploded' burger, showing an assembled one instead.
• − Contains a spelling error in the word 'ONLY' ('LNHLY').

GPT Image 1

• + Successfully rendered the 'exploded' burger with components suspended in mid-air.
• + Captured the 'fiery, glowing effect' on the text much better than the competitor.
• + High photorealistic detail on the ingredients, especially the patty and tomatoes.

• − Missing the number '6' in the price starburst (renders as '€.99').
• − The composition feels a bit tight at the top, cutting off the very edge of the title.

FLUX.1 Kontext [dev]

• + Excellent authentic chalk texture with realistic smudging and pressure variations.
• + Great overall composition including the wooden frame and warm café lighting.
• + High contrast makes it look like a real physical object.

• − Significant spelling errors ('Risoktso', 'Mashroom', 'Octpus').
• − Text repetition and overlapping artifacts in the pricing section.
• − Failed to follow the date formatting accurately, using nonsense characters for the month.

GPT Image 1

• + Perfect text rendering with zero spelling mistakes.
• + Accurately followed the date instructions and provided the full menu items requested.
• + Very clean and readable layout.

• − The font looks slightly too uniform, more like a digital 'chalk' font than natural handwriting.
• − Lacks the realistic chalk dust and grain texture found in the first image.
• − The framing is cropped too tightly compared to the ambient view of the first model.

FLUX.1 Kontext [dev]

• + Successfully followed the difficult logic of the horse riding the astronaut.
• + High facial detail and clean rendering of the space suit.
• + Maintains a clean, modern cinematic aesthetic.

• − The 'riding' stance is a bit ambiguous as the horse is floating behind and slightly above.
• − The stars/planets in the background look like simple gray spheres.

GPT Image 1

• + Excellent atmospheric lighting and texture.
• + Dynamic composition with a powerful sense of movement.

• − Completely failed the semantic instruction, showing an astronaut riding a horse instead of the reverse.
• − Cliché interpretation that ignores the specific spatial prompt.

FLUX.1 Kontext [dev]

• + Excellent fur texture and lighting on the capybara
• + High detail in the woman's face and hands
• + Balanced composition with good depth of field

• − The capybara only has one paw on the wheel instead of two as requested
• − The capybara's face looks slightly more like a groundhog or rodent mix than a pure capybara

GPT Image 1

• + Followed the instruction for both front paws on the steering wheel
• + Very realistic capybara facial anatomy
• + More iconic 'Taxi' hat design that is clear and legible

• − The woman in the background is quite blurry compared to Image A
• − The capybara's left hand/paw on the wheel has slightly strange, finger-like anatomy

FLUX.1 Kontext [dev]

• + Strong, high-contrast lighting on the central jack-o-lantern.
• + Clean and legible header text.

• − Garbled and illegible text within the scroll banner.
• − Incorrect spelling of the location 'The Arches' and cluttered date/time line.
• − Flat background that lacks the requested 'moody night sky' or parchment texture.

GPT Image 1

• + Perfectly accurate spelling for the scroll banner and main title.
• + Superior gothic aesthetic with dark parchment texture and moody atmospheric lighting.
• + Incorporates all requested elements including webs, thorns, and a moonlit sky.

• − Swapped the location and time labels, putting the address next to the word 'Time'.

FLUX.1 Kontext [dev]

• + Natural-looking hair texture and shine.
• + Excellent preservation of the background and clothing.
• + The hairstyle looks modern and well-groomed.

• − Significantly altered the facial features and glasses, making the person look younger and different from the source.
• − Skin texture on the forehead was smoothed out excessively.

GPT Image 1

• + Very high preservation of the original facial features, skin texture, and glasses.
• + Added hair that matches the messy, rugged aesthetic of the source image.
• + Maintains the original lighting and color grading perfectly.

• − The hair volume is slightly exaggerated, giving it a wig-like appearance at the top.
• − Hair texture is a bit coarse compared to the beard.

FLUX.1 Kontext [dev]

• + Extremely clean and bold typography
• + High-clarity renders with very soft, smooth textures
• + Perfectly followed the 'solid light blue background' instruction

• − The flag icon is stylized and technically incorrect
• − The sushi anatomy is slightly nonsensical with the black band wrapping around the rice and fish horizontally

GPT Image 1

• + Excellent adherence to all prompt elements including the flag and text hierarchy
• + High-quality 3D miniature aesthetic with realistic clay-like PBR textures
• + Beautiful composition with a logical variety of sushi and garnish

• − The background is a slightly darker, muted blue rather than a bright light blue
• − Minor texture noise on the diorama base compared to Model A's extreme smoothness

FLUX.1 Kontext [dev]

• + Successfully captures the subject's clothing and pose from the original photo in a comic style.
• + Inclusion of a TV showing the anchor persona.
• + Clean graphic illustration style.

• − Completely missed the 'hockey' requirement.
• − The 'JOB' text is redundant and lacks creativity.
• − The dog is more of a generic clip-art style rather than a humorous caricature.

GPT Image 1

• + Excellent adherence to all prompts: news anchor desk, dog, and hockey elements are all present.
• + High artistic quality with a classic watercolor caricature aesthetic.
• + Highly creative integration of the dog playing hockey on the TV screen.

• − The facial likeness is somewhat less accurate to the original subject compared to model A.
• − Slight anatomical issues on the hands holding the papers.

FLUX.1 Kontext [dev]

• + Excellent clarity and sharpness in the foreground
• + Vibrant colors and high-contrast lighting

• − Failed to include the fox and bunny which were explicitly requested
• − Anatomical issues with the central kitten's front paws
• − Overall composition feels a bit digital and less natural

GPT Image 1

• + Successfully included all four requested animals (dog, cat, bunny, fox)
• + Outstanding atmospheric lighting with god rays and dew sparkles as requested
• + More realistic fur textures and higher detail in facial expressions

• − The fox kit has six legs/paws visible due to a generation artifact
• − The kitten has an oddly long, stretched body where it connects to its back paws

FLUX.1 Kontext [dev]

• + Excellent preservation of the original image's composition and clothing patterns
• + Clean lines and sharp details
• + Accurately replicates the characters' facial expressions and poses from the meme

• − Art style leans more towards generic modern anime than the specific Ghibli aesthetic
• − Lighting is flat and lacks the 'dreamy' quality requested

GPT Image 1

• + Strong adherence to the Ghibli-inspired request with soft pastel colors and hand-painted textures
• + Captured the warm, nostalgic mood and dreamy lighting perfectly
• + Excellent artistic interpretation of the scene

• − Loses the specific plaid pattern on the man's shirt
• − Character faces are less recognizable as the individuals from the original source photo

FLUX.1 Kontext [dev]

• + Excellent source preservation, maintaining the identity and environment almost perfectly.
• + Subtle, natural-looking hair motion that flows realistically from the head.

• − Minimal leaves added, which makes the 'flying leaves' part of the prompt feel under-delivered.
• − The dynamic motion effect is very subtle compared to the request for an 'energetic' feel.

GPT Image 1

• + Strong adherence to all parts of the prompt, including significant hair motion and numerous flying leaves.
• + Successfully conveys a much more 'energetic and lively' atmosphere as requested.

• − Noticeable loss of source preservation; the woman's facial features and the dog's face have been altered.
• − Some visual artifacts where flying leaves overlap with the characters.

FLUX.1 Kontext [dev]

• + Excellent typography rendering with the correct accent on 'Caffè'.
• + Clean, minimalist vector aesthetic.
• + Perfectly balanced layout and vertical alignment.

• − The dome interpretation looks more like a circus tent or architecture than a food cloche.
• − Missing the requested banner element for the 'Est. 1720' text.
• − Very little visible texture on the background.

GPT Image 1

• + Successfully includes all prompt elements including the banner and steam.
• + Beautiful grainy texture that accurately reflects the 'vintage' and 'subtle texture' request.
• + Classic serif typography fits the 1720 era well.

• − The cloche dome icon is slightly asymmetrical.
• − Failed the 'light background' instruction by using a black background.

FLUX.1 Kontext [dev]

• + Stronger adherence to the navy and muted red color palette.
• + Includes 'Apollo 11' text prominently.

• − Icons are messy, abstract, and do not clearly represent the requested steps like a Saturn V or Lunar Module.
• − Text rendering is poor with significant gibberish and overlapping characters.
• − Layout is chaotic and fails as a functional infographic.

GPT Image 1

• + Successfully includes clear, recognizable icons for Saturn V, Earth Orbit, and the Lunar Module.
• + Provides much better text legibility and logical structure for an infographic.
• + Includes creative and accurate supporting details like the crew names and 'Tranquility' location.

• − Slightly deviates from requested steps by merging/renaming steps (e.g., 'Earllunar').
• − The composition feels a bit cramped with the white text bars.