GPT Image 1 OpenAI vs OmniGen v2 VectorSpaceLab

GPT Image 1

• + Excellent adherence to the 'partially visible through the glass' instruction for the plant.
• + Highly realistic glass textures and refractions around the edges.
• + Accurate lighting direction and soft shadows.

• − The glass cube appears to have an open front or missing face rather than being a solid closed object.

OmniGen v2

• + Successfully includes all requested elements in a clean composition.
• + Beautiful glossy texture on the blue sphere with realistic reflections.
• + Sharp focus on the central subject.

• − The plant is positioned entirely above the cube, missing the requested effect of being visible through the glass.
• − The glass cube looks a bit more like a thick acrylic box than glass.

GPT Image 1

• + Excellent skin texture and realistic, aged features on the subject.
• + Captures the requested 'imperfect framing' and 50mm feel perfectly.
• + Authentic atmosphere with subtle rain, wet textures, and natural lighting.

• − The rear bicycle structure is a bit anatomically messy around the chain guard area.
• − Missing significant motion blur from the passing cars as requested.

OmniGen v2

• + Strong reflections on the wet pavement.
• + Clearer representation of a full bicycle.

• − Skin texture appears smoothed and lacks the 'natural skin texture' requested.
• − Composition feels like a staged studio portrait rather than a 'candid street photo'.
• − Rain effect looks like a simple digital overlay of streaks rather than environmental moisture.

GPT Image 1

• + Excellent depiction of a battle-worn character with realistic skin texture and grit
• + Highly detailed and realistic engraving on the plate armor
• + Atmospheric lighting that feels natural to the environment

• − The beads in the hair are very dark and blend into the shadows

OmniGen v2

• + Clearly visible beads in the braided hair as requested
• + Clean and vibrant bokeh effect in the background
• + Good focus on the lifelike quality of the eyes

• − The character looks too clean and polished for a 'battle-worn' description
• − Dirt marks look like flat digital face paint rather than actual grime
• − Armor texture looks slightly plastic compared to Model A

GPT Image 1

• + Excellent typography and readability with clean sans-serif fonts
• + High-quality, appetizing food photography that looks professional
• + Clear layout with relevant sections that match the prompt

• − The placeholder text contains spelling errors like 'descrigion'
• − The grid is a bit tight with limited white space between the photos and text

OmniGen v2

• + Successfully creates a double-page spread menu layout
• + Uses vibrant color blocks and accents as requested
• + Better representation of a 'grid' layout for the overall design

• − Text is largely nonsensical gibberish or has severe spelling errors
• − The resolution of the food photos is lower compared to Model A
• − The fonts used are more stylistic/rustic rather than a 'bold sans-serif'

GPT Image 1

• + Excellent photorealistic texture on the meat and bun
• + Successfully rendered all requested text elements with the corect font style
• + Captures the 'exploded' suspended animation look perfectly

• − The price text is missing the leading '6' in the starburst

OmniGen v2

• + Bold, high-contrast graphic design suitable for a cartoonish ad
• + Corretly includes the full '6.99' numerical value in the starburst

• − Failed the 'exploded burger' requirement; the burger is fully assembled
• − Text is partially cut off on the left side
• − Lacks the photorealism requested in the prompt, appearing more like a 3D render

GPT Image 1

• + Excellent text legibility and spelling throughout the entire menu.
• + Realistic chalk texture with convincing grainy artifacts.
• + Consistent handwriting style that feels authentic to a café setting.

• − The title is in a print-style sans-serif rather than the requested 'elegant cursive'.
• − The layout is a bit sparse with significant empty space at the sides.

OmniGen v2

• + Includes a wooden frame which adds to the 'cozy café' atmosphere.
• + Attempts a more varied handwriting style for the title and body text.

• − Numerous spelling errors and gibberish text throughout the list.
• − Poorly rendered letterforms with overlapping lines and messy corrections.
• − Failed to follow the specific menu item text precisely.

GPT Image 1

• + Excellent cinematic lighting and textures.
• + Highly detailed rendering of the spacesuit and horse anatomy.

• − Failed the core prompt instruction to have the horse on top of the astronaut.

OmniGen v2

• + Clean, vibrant colors.
• + High contrast and sharp lines.

• − Failed the core prompt instruction to have the horse on top of the astronaut.
• − Visual style is more illustrative than 'cinematic' or 'surreal'.

GPT Image 1

• + Excellent photorealism in texture and lighting
• + Captures the professional, calm expression of the capybara perfectly
• + Accurate depiction of a capybara's anatomy for the paws on the wheel

• − The view is from outside looking in, while the prompt requested a scene 'inside'

OmniGen v2

• + Dynamic composition and angle
• + Good interpretation of the driver cap

• − Anatomical failure where human hands are growing out of the capybara's arms
• − Perspective issues where the capybara seems to be floating in front of the seat
• − The businesswoman is in the front passenger seat rather than the requested back seat

GPT Image 1

• + Excellent typography with perfect spelling of all requested text.
• + Sophisticated, authentic vintage gothic aesthetic with moody lighting.
• + Clear and balanced composition that functions well as a real invitation.

• − The parchment texture is a bit dark, making the tree silhouettes blend into the background.

OmniGen v2

• + High contrast lighting makes the jack-o-lantern and moon pop.
• + Creative use of a deckled-edge parchment border effect.

• − Significant spelling errors and garbled text on the banner and bottom details.
• − The visual style is more like a modern digital illustration than a vintage gothic poster.
• − Poor layout of event details, resulting in a cluttered and confusing bottom section.

GPT Image 1

• + Excellent typography with clean, bold text and a correct Japanese flag icon.
• + Higher quality 3D textures with realistic soft-touch PBR materials.
• + Superior object modeling, especially in the rice grains and sushi garnish.

• − The perspective is slightly more focused/closer than a true 'miniature' diorama look.

OmniGen v2

• + Stronger adherence to the 'isometric' diorama layout with a distinct raised base.
• + Clean, bright color palette that pops against the blue background.

• − Incorrect flag icon that does not resemble the Japanese flag.
• − Visible AI artifacts in the sushi details, such as the shrimp tail merging into nigiri.
• − Text has slight rendering errors and inconsistent shadows.

GPT Image 1

• + Excellent adherence to the 'hyper-photorealistic' instruction with realistic fur textures and lighting.
• + Correctly includes all four requested animals: puppy, kitten, bunny, and fox kit.
• + Dynamic composition that captures the 'tumbling' and 'chasing' actions described.

• − One butterfly has slightly clipped wings in the upper left.
• − The kitten's anatomy is a bit merged with the puppy's leg area.

OmniGen v2

• + Bright, vibrant colors that fit a 'wholesome' vibe.
• + Clear, centered composition suitable for a greeting card.

• − Failed to include the bunny, only showing three animals.
• − Style is closer to 3D animation/clipart rather than the requested 'hyper-photorealistic' look.
• − The animals are sitting still rather than 'chasing' or 'tumbling'.

GPT Image 1

• + Excellent adherence to the 'hand-painted texture' and 'dreamy' lighting request
• + Perfectly captures the Studio Ghibli watercolor aesthetic
• + Preserves the characters' expressions and the specific composition of the original meme.

• − Image looks slightly washed out compared to modern digital anime styles.

OmniGen v2

• + Clean, high-resolution digital illustration style
• + Preserves details like the plaid pattern on the shirt and the background bus very well.

• − Fails to capture the 'Studio Ghibli' look, opting for a generic modern anime style
• − Loses the key narrative of the meme by making the girlfriend look happy instead of angry
• − Does not follow the request for soft pastel colors or hand-painted textures.

GPT Image 1

• + Excellent source preservation, maintaining the woman's face and dog's appearance perfectly.
• + Logical hair movement that flows naturally from the head.
• + Leaves appear integrated into the original scene's lighting.

• − The wind effect on the dog's fur is slightly subtler than requested.

OmniGen v2

• + Successfully adds a clear wind-blown effect to the hair.
• + Leaves are bright and highly visible.

• − Failed to preserve the woman's facial identity, significantly changing her appearance.
• − Leaves look like stickers or icons rather than natural elements of the scene.
• − Noticeable loss of detail and texture in the dog's fur compared to the source.

GPT Image 1

• + Excellent typography with perfect spelling and correct accent marks.
• + Clean vector-style execution with a subtle, professional texture.
• + Balanced layout that feels authentic to the vintage prompt.

• − Ignored the 'light background' instruction, delivering a black background.

OmniGen v2

• + Followed the light/cream background instruction correctly.
• + Good use of the banner element for the 'Est. 1720' text.
• + Clean, minimalist line work on the cloche.

• − Major spelling error in the main brand name ('CAFFFLORIN').
• − Typographic hierarchy is slightly cluttered with overlapping elements.
• − Steam effect is very thin and lacks visual weight compared to the rest of the logo.

GPT Image 1

• + Excellent text rendering with names and phases correctly spelled
• + Followed the flat-vector style and specific NASA color palette perfectly
• + Logical layout that visualizes the progression of the mission mission phases

• − Included a typo in the bottom bar ('EARLLUNAR')
• − Disconnected 'Lunar Orbit' and 'Descent' icons from their labels

OmniGen v2

• + Clean icon designs within circular frames
• + Modern, bold typographic header

• − Extremely poor text rendering with numerous gibberish words and spelling errors
• − Incorrect mission number (Apollo 17 instead of 11)
• − Failed to follow the specific 6-step sequence requested in the prompt