GPT Image 1 Mini OpenAI vs OmniGen v2 VectorSpaceLab

GPT Image 1 Mini

• + Excellent photographic realism with natural textures on the book and table.
• + Perfect adherence to the lighting instruction with a soft glow from the left.
• + Accurate glass refraction and transparency showing the plant behind.

• − The glass cube looks more like a frame or open box due to the very thin edges.

OmniGen v2

• + Strong prompt adherence for all spatial relationships.
• + The glass cube has realistic thickness and weight.
• + Vibrant colors and clean composition.

• − The red book appears to be clipping into or through the top of the glass cube.
• − Lighting is a bit more generic and less 'soft window light' than model_a.
• − The table surface looks slightly more CGI than natural wood.

GPT Image 1 Mini

• + Excellent photographic realism with natural skin texture and fine details like raindrops and worn metal.
• + Authentic cinematic lighting and color grading that matches the 'candid' and 'no stylization' request.
• + Perfect adherence to the 50mm shallow depth of field and motion-blurred background cars.

• − The 'imperfect framing' is very subtle, bordering on a professional tight crop.

OmniGen v2

• + Strong reflections on the wet pavement.
• + Clear depiction of the red bicycle color requested.

• − The image looks highly stylized and artificial, failing the 'no stylization' and 'realistic' prompt.
• − The 'repairing' action is not shown; the man is just standing with the bike.
• − Visual artifacts present, such as the bike pedals merging with the man's leg and the floating bike stand.

GPT Image 1 Mini

• + Excellent depiction of battle-worn texture and grit on the face and armor.
• + Highly detailed engraving on the plate armor with realistic lighting reflections.
• + Captures a serious, moody atmosphere consistent with a battle-hardened character.

• − The braids are very thin and subtle, lacking the clear 'small beads' requested in the prompt.

OmniGen v2

• + Successfully includes the requested beads within the hair braids.
• + Good warm lighting from the background torch/fire source.
• + Clean composition with a clear shallow depth of field.

• − The character looks too clean and 'model-like' for someone described as 'battle-worn'.
• − The dirt/scars look like superficial spots rather than realistic skin texture.
• − The armor engraving is less detailed and looks more generic compared to Model A.

GPT Image 1 Mini

• + Excellent text rendering with no spelling errors
• + Clean, logical grid layout that aligns with the requested sections
• + High-quality, realistic food photography

• − The design is a bit sparse with no placeholder item text under the headers

OmniGen v2

• + Includes realistic placeholder text blocks and a more complete menu layout
• + Uses vibrant colored accents as requested in the prompt

• − Multiple spelling errors in the headers like 'Restaurated Ments' and 'Apptetizes'
• − The food categories do not match the prompt requests specifically
• − Composition feels a bit cluttered compared to a minimalist style

GPT Image 1 Mini

• + Excellent adherence to the 'exploded' and 'suspended' burger request.
• + Flawless text rendering for all three required text elements.
• + High photorealistic detail in the textures of the meat and bun.

• − The fiery embers in the background are a bit subtle compared to the request for 'fiery background'.

OmniGen v2

• + Strong 'fiery' glow effect behind the main subject.
• + Clean, professional-looking graphic design layout.

• − Failed the 'exploded' burger instruction, showing a standard assembled burger.
• − Text is cut off ('LIMITED TIM...') and is missing the requested fiery glow effect.
• − Missing the Euro symbol (€) in the pricing starburst.

GPT Image 1 Mini

• + Excellent text rendering with near-perfect spelling across all items.
• + Captures a realistic chalk texture with grainy edges and natural board artifacts.
• + Follows the layout instructions faithfully, including the logical completion of the truncated prompt 'Brown But...'.

• − The title is in all-caps rather than the requested elegant cursive style.
• − The handwriting is a bit too uniform, bordering on looking like a digital font despite the chalk texture.

OmniGen v2

• + Higher contrast and bolder chalk presentation.
• + The frame has a clean appearance.

• − Numerous spelling errors including 'SPECALS', 'Musonhom', 'Lemont', and 'cllip'.
• − Severe text jumbling and overlapping characters in the middle section.
• − Failed to maintain a consistent style, with the title appearing much more 'digital' than the disorganized body text.

GPT Image 1 Mini

• + Excellent cinematic lighting and textures.
• + Highly detailed and realistic rendering of the horse's musculature and the spacesuit.
• + Sophisticated composition with a natural sense of motion.

• − Failed the specific spatial instruction; the astronaut is on top, not the horse.

OmniGen v2

• + Clean, vibrant colors.
• + Correctly placed in a space-like background.

• − Failed the specific spatial instruction; the astronaut is riding the horse.
• − Visual style is somewhat generic and lacks the requested cinematic detail.
• − Floating moons look disjointed and lack depth.

GPT Image 1 Mini

• + Excellent photorealistic texture on the capybara's fur and the jacket.
• + Cinematic lighting and atmospheric bokeh that truly feels like New York at night.
• + Accurately depicts the capybara's paws on the steering wheel.

• − The passenger is very blurry, though this adds to the depth of field effect.
• − Only one paw is clearly visible on the wheel despite the prompt asking for both.

OmniGen v2

• + Captures a higher level of brightness and clarity in the vehicle interior.
• + Successfully includes two hands/paws on the steering wheel.

• − Serious anatomical failure where human hands are attached to the capybara's arms.
• − The composition is confusing, as the passenger appears to be sitting in the front passenger seat rather than the back seat.
• − The image has a more synthetic, CGI look compared to the requested photorealism.

GPT Image 1 Mini

• + Excellent typography with perfect spelling and clear legibility.
• + Sophisticated, moody atmosphere that perfectly matches the 'vintage gothic' and 'dark parchment' description.
• + High visual coherence with a professional, balanced layout.

• − The parchment texture is a bit subtle across the whole image compared to a defined paper edge.

OmniGen v2

• + Strong 'parchment' effect with torn edges.
• + Vibrant colors and high-contrast lighting.

• − Numerous spelling errors including 'FRIGTS' and 'THE ARCAS'.
• − Poor text rendering on the scroll banner which contains illegible symbols.
• − The composition feels cluttered with overlapping and repetitive text elements.

GPT Image 1 Mini

• + Excellent text rendering and typography layout.
• + High-quality soft 3D textures that accurately reflect the 3D cartoon request.
• + Accurate Japanese flag icon and realistic sushi varieties.

• − Perspective is slightly flatter than a strict 45° isometric angle.

OmniGen v2

• + Perfect 45° isometric projection on the diorama base.
• + Bold, high-contrast colors that pop well.
• + Good composition of the 3D diorama block.

• − Incorrect flag representation (unknown yellow/red flag).
• − Unusual black square artifact inside the sushi rolls.
• − Inconsistent text lighting with shadows that don't match the scene.

GPT Image 1 Mini

• + Excellent adherence to the 'hyper-photorealistic' part of the prompt with realistic fur textures and anatomy.
• + Captures the action of 'chasing and tumbling' with dynamic, mid-air poses.
• + Includes all four requested animals: puppy, kitten, bunny, and fox kit.

• − The transition from the puppy's back to the background lighting is slightly soft.
• − The second butterfly has a somewhat simplistic wing shape.

OmniGen v2

• + Bright, vibrant colors that evoke a joyful vibe.
• + Clear, distinct butterfly subjects positioned well in the frame.

• − Failed the photorealism requirement, producing a cartoonish or 3D-render aesthetic instead.
• − Missed one of the four requested animals, showing only three.
• − The animals are sitting statically rather than 'chasing and tumbling' as requested.

GPT Image 1 Mini

• + Excellent preservation of the original facial expressions and subtle nuances of the 'Distracted Boyfriend' meme.
• + Beautiful painterly texture that feels like colored pencil or soft pastel on paper.
• + Warm, nostalgic color palette that perfectly fits the dreamy mood requested.

• − The style is more 'classic illustration' than the specific clean-line anime look associated with modern Ghibli films.

OmniGen v2

• + Successfully adopts a clean, cel-shaded anime aesthetic common in Japanese animation.
• + Preserves the composition and clothing colors effectively.

• − Fails to preserve the critical facial expressions, making the girlfriend look happy/neutral rather than shocked/angry.
• − Lacks the soft, hand-painted textures and 'dreamy' lighting requested in the prompt.
• − The background architectural style feels generic and modern rather than Ghibli-esque.

GPT Image 1 Mini

• + Successfully added blowing hair and flying leaves.
• + Maintained high facial similarity to the source image.
• + Preserved the background elements like the bridge more accurately.

• − The leaf rendering is a bit blurry and inconsistent in scale.
• − Slight alteration to the lighting of the scene compared to the original.

OmniGen v2

• + Excellent hair motion effect with realistic strands.
• + Clean, vibrant visual quality and high contrast.
• + Good placement of flying leaves that integrates well with the sky.

• − Significantly changed the facial features of the woman.
• − Leaves look somewhat synthetic and uniform in color.
• − The dog's face is slightly altered from the source image.

GPT Image 1 Mini

• + Excellent typography including the grave accent in 'Caffè'
• + Precise rendering of all requested text including 'Est. 1720'
• + Rich texture and high-quality vintage aesthetic

• − Failed to provide a light background as requested in the prompt

OmniGen v2

• + Followed the light background and warm brown/cream tone instructions well
• + Strong minimalist vector style
• + Clean and balanced composition

• − Misspelled the primary brand name as 'CAFFFLORIN'
• − Graphic for the cloche is slightly abstract and lacks detail

GPT Image 1 Mini

• + Excellent text rendering with no spelling errors.
• + Perfect adherence to all six requested steps in the correct order.
• + Clean, professional flat-vector aesthetic with consistent iconography.

• − The 'Translunar' icon is a bit abstract and messy compared to the others.
• − The layout is slightly cramped at the bottom.

OmniGen v2

• + Strong 'modern infographic' layout with distinct sections.
• + Good use of the requested NASA-inspired color palette.

• − Severely garbled text and spelling errors (e.g., 'APOLO 17', 'NSA').
• − Failed to include the specific six steps requested in the prompt.
• − Iconography is inconsistent and does not match the prompt descriptions.