FLUX1.1 [pro] Black Forest Labs vs GPT Image 2 OpenAI

FLUX1.1 [pro]

• + Excellent photorealistic lighting and window reflections on the glass and sphere.
• + Highly realistic texture on the wooden table and red book cover.
• + Sharp details and high visual clarity.

• − The glass container is a tall rectangular prism rather than a cube as requested.
• − The sphere appears to be levitating unnaturally rather than resting.

GPT Image 2

• + Perfect adherence to the geometric 'cube' shape requested.
• + Natural placement of the small sphere resting on the bottom of the cube.
• + Soft, realistic lighting that accurately captures the window light direction.

• − The texture of the book and plant is slightly softer and less detailed than Image A.
• − The glass edges look a bit thick, resembling a fish tank more than a solid glass cube.

FLUX1.1 [pro]

• + Excellent cinematic atmosphere and lighting
• + Captures a gritty urban feel with beautiful wet pavement reflections
• + Follows the motion blur and framing instructions well

• − The anatomy of the bicycle is mangled and physically impossible
• − The man is not actually 'repairing' the bike, just standing over it
• − Heavy artifacts around the man's feet and the bike frame

GPT Image 2

• + Highly realistic bicycle and tool set details
• + Clearer adherence to the 'repairing' action with appropriate tools and posture
• + Features readable and contextually accurate Japanese text on the sign

• − Does not capture the 'motion blur' requested in the prompt
• − The depth of field feels less like a 50mm lens and more like a standard smartphone photo
• − The lighting is flat compared to the cinematic request

FLUX1.1 [pro]

• + Extremely high-fidelity skin texture with realistic pores and sweat details.
• + Excellent execution of bokeh sparks and dramatic torchlight contrast.
• + Captures an intense, battle-worn expression that fits the 'paladin' theme well.

• − The braids are very subtle and located mostly in the background hair, making them difficult to see.
• − The armor is less prominent due to the extreme close-up crop.

GPT Image 2

• + Excellent adherence to the 'hair braided with small beads' part of the prompt.
• + Displays beautiful, intricate engraving on the plate armor and visible leather straps.
• + Effective use of warm lighting and shallow depth of field for a cinematic look.

• − The character looks relatively clean compared to the 'battle-worn' and 'faint scars' request.
• − The sparks / bokeh elements are much less prominent than in the other image.

FLUX1.1 [pro]

• + Excellent presentation of the menu in a real-world context with silverware
• + Maintains a clean, minimalist aesthetic with plenty of white space

• − Text is largely illegible gibberish
• − Food photos are small and lack specific detail

GPT Image 2

• + Perfect text rendering for all menu items and descriptions
• + High-quality, distinct food photography in a clear grid
• + Strict adherence to all prompt sections (appetizers, pizza, mains)

• − Slightly less 'minimalist' than a high-end editorial design due to dense information

FLUX1.1 [pro]

• + Crisp and clean text rendering.
• + Good photorealistic texture on the meat and buns.

• − Failed to provide an 'exploded' burger, showing a mostly assembled one instead.
• − Included blue neon text which contradicts the 'fiery, glowing effect' request.
• − Failed to include the starburst for the price.

GPT Image 2

• + Perfect adherence to the 'exploded' burger concept with suspended components.
• + Excellent fiery, glowing text effects applied to all requested text elements.
• + Strictly followed the starburst requirement for the price.

• − The composition is slightly more cluttered than Model A.
• − Minor artifact where some sauce droplets look a bit plastic-like.

FLUX1.1 [pro]

• + Excellent background blurring and café atmosphere
• + Clean, legible handwriting style

• − Confused the octopus menu item and price, listing it twice with errors ('$228' and 'Herbss')
• − Text rendering looks more like a digital font than actual chalk on board
• − Spelling error in 'Chipkies' and 'giluten'

GPT Image 2

• + Perfect adherence to the complex list of menu items and prices
• + Very realistic chalk texture with dusty smudges and varying pressure
• + Accurate handwriting variations as requested in the prompt

• − The frame of the chalkboard is slightly cropped at the bottom
• − Composition is a bit tighter than Model A

FLUX1.1 [pro]

• + High cinematic visual quality and lighting
• + Detailed space environment
• + Coherent horse anatomy and textures

• − Failed the negative constraint: the astronaut is riding the horse instead of the horse on top

GPT Image 2

• + Perfect adherence to the complex spatial constraint of horse over astronaut
• + Surreal and humorous interpretation of the prompt
• + Clean texture rendering on the spacesuit and horse fur

• − The horse's front legs terminate awkwardly into the reins
• − The anatomy of the astronaut as a four-legged mount is slightly distorted

FLUX1.1 [pro]

• + Excellent photographic lighting and cinematic bokeh.
• + High resolution and very clean textures on the capybara fur.
• + Well-rendered side profile of the businesswoman.

• − Compositional error: the woman appears to be sitting in the front passenger seat next to the driver, not the back seat.
• − The capybara's paws are not on the steering wheel as requested.
• − The capybara looks too small or incorrectly positioned relative to the seat.

GPT Image 2

• + Accurately places the businesswoman in the back seat as requested.
• + Successfully shows the capybara's paws on the steering wheel.
• + The capybara's proportions and placement in the driver's seat are much more realistic.

• − The image quality is slightly softer with less sharp detail compared to the other model.
• − The cap looks a bit like a police hat rather than a standard yellow taxi cap.
• − Minor blurring on the woman's face and hands.

FLUX1.1 [pro]

• + Excellent character rendering on the jack-o-lantern
• + Clean thorn border artwork
• + Atmospheric use of the moon and silhouettes

• − Significant text repetition and spelling errors in the invitation details
• − Missed several text elements from the prompt
• − Large empty black space at the bottom feels unbalanced

GPT Image 2

• + Perfect adherence to all text requirements with 100% spelling accuracy
• + Rich, detailed gothic composition with webs, thorns, and integrated NYC skyline
• + Followed the square format request perfectly while maintaining a vintage parchment feel

• − The jack-o-lantern is a bit dark compared to the 'glowing' request
• − The border is very busy, though it fits the requested theme

FLUX1.1 [pro]

• + Excellent soft 3D cartoon style consistent with the prompt
• + Beautifully integrated soft lighting and refined textures
• + Perfectly matches the 'minimal garnish' and clean background aesthetic

• − Missed the small flag icon requested in the prompt
• − Text for 'SUSHI' is relatively thin and lacks professional hierarchy

GPT Image 2

• + Successfully included all prompt elements including the small flag icon
• + Outstanding text rendering with bold, clean typography
• + Very intricate textures on the diorama base and food items

• − Less of a 'cartoon' feel and more of a realistic CGI render
• − Diorama is much busier than the requested 'minimal' design

FLUX1.1 [pro]

• + Excellent whimsical lighting and backlighting effects on fur.
• + Very high artistic consistency and cute factor.
• + Strong focus on the 'expressive eyes' part of the prompt.

• − Failed to include the red fox kit.
• − The animals are mostly sitting rather than 'tumbling' or 'chasing'.
• − The butterflies are glowing blobs rather than detailed insects.

GPT Image 2

• + Successfully included all four requested animals: puppy, kitten, bunny, and fox.
• + Better adherence to the dynamic action of 'chasing and tumbling'.
• + Features clear god rays and more detailed butterfly renderings.

• − The fox kit has somewhat unnatural black legs/paws that look like socks.
• − The kittens face is slightly distorted in its anatomy.
• − Background wildflowers are a bit cluttered compared to the subject focus.

FLUX1.1 [pro]

• + Excellent vector emblem style that feels cohesive and professional.
• + Sophisticated use of warm brown and cream tones.
• + Strong decorative flourishes that match a vintage aesthetic.

• − Failed the primary text spelling by rendering 'CAFFÈ FRANILIAN' instead of 'Caffè Florian'.
• − The dome structure looks more like a building cupola than a food cloche.

GPT Image 2

• + Perfect adherence to text, including proper spelling and accents.
• + Accurately depicts a food cloche dome as requested.
• + Exceptional subtle texture on the background and logo elements.

• − The 'EST. 1720' text is slightly less crisp than the main brand name.
• − The steam lines are a bit more illustrative and less 'minimalist' vector-style than Model A.

FLUX1.1 [pro]

• + Features a minimalist artistic style with a large, central rocket graphic.
• + Uses the requested NASA-inspired color palette effectively.

• − Text is largely illegible gibberish and full of spelling errors (e.g., 'Earna Orbit', 'Desceing').
• − The logical flow of information is disorganized, numbering from 1 up to multiple 6s and placing steps out of order.
• − The icons do not match the specific requested items accurately.

GPT Image 2

• + Excellent adherence to all six requested infographic steps with accurate icons.
• + Perfect text rendering for headers, steps, and crew names.
• + Clean, professional layout that genuinely resembles a modern educational poster.

• − The 'Descent' and 'Landing' modules are slightly over-detailed relative to a strictly 'flat-vector' style.
• − The Saturn V rocket shows a small amount of photo-realistic smoke at the base which breaks the vector aesthetic.