GPT Image 1 Mini OpenAI vs Grok Imagine Image Pro xAI

GPT Image 1 Mini

• + Excellent photographic realism and texture in the book and sphere.
• + Accurate lighting direction consistent with the prompt.
• + Perfectly clear glass rendering with subtle reflections.

• − The blue sphere is floating or partially intersecting the back glass pane rather than sitting on the surface.
• − The cube structure has an odd, thin metal-like frame rather than being solid glass.

Grok Imagine Image Pro

• + Physically grounded sphere sitting on the base of the cube.
• + Dynamic refraction of the sphere and plant within the thick glass.

• − The glass geometry is warped and inconsistent at the corners.
• − The blue sphere is inexplicably reflected/duplicated on the right side within the glass despite no physical wall being there.

GPT Image 1 Mini

• + Excellent shallow depth of field and bokeh transition
• + Strong adherence to 'imperfect framing' with a close-up, candid feel
• + Atmospheric lighting and realistic ground reflections

• − Anatomical issues with the man's hands merging into the bike spokes
• − Missing visible motion blur from passing cars
• − The bicycle has structural issues where the kickstand and frame meet

Grok Imagine Image Pro

• + Successfully includes motion blur from passing cars
• + Better anatomical correctness in the hands and clear use of a tool
• + Environment feels more like a specific Japanese street scene

• − Depth of field is slightly too deep compared to the 50mm request
• − Street reflections are less pronounced than in Model A
• − The image looks a bit more 'clean' and less like a moody candid shot

GPT Image 1 Mini

• + Excellent photographic realism in skin texture and eyes
• + Naturalistic lighting with subtle warm hues
• + Very detailed engraving on the plate armor

• − Failed to include the requested beads in the braided hair
• − Scars are very faint, leaning more towards just dirt

Grok Imagine Image Pro

• + Followed all prompt instructions including hair beads and distinct scars
• + Impressive detail on leather straps and cloth underlayer
• + Dynamic composition with clear Latin text engraving

• − The sparks look like flat digital streaks rather than bokeh light
• − Skin texture looks slightly more airbrushed compared to Model A

GPT Image 1 Mini

• + Excellent chalk texture with grainy, realistic strokes
• + Perfectly centered and clean composition
• + High-fidelity rendering of the wooden frame

• − Failed to provide 'elegant cursive' for the title as requested
• − Text looks a bit too uniform, bordering on a digital font appearance

Grok Imagine Image Pro

• + Successfully used cursive for the menu items to create a more authentic handwritten feel
• + Captures the 'cozy café' atmosphere with a visible background and environmental lighting
• + Variation in handwriting styles makes it look more realistic

• − The title is in print instead of the requested 'elegant cursive'
• − Text layout is slightly less balanced with some spacing issues near the prices

GPT Image 1 Mini

• + Successfully transferred the character's facial features, sunglasses, and clothing style
• + Accurately matched the lighting and background color of the first source image
• + Correctly interpreted the identity transformation requested

• − Significantly simplified and altered the specific complex pose from Image 1
• − Anatomical issues in the right hand and feet placement compared to the original pose

Grok Imagine Image Pro

• + Maintained the exact complex pose from Image 1 with high fidelity
• + High visual quality and resolution of the overall image

• − Completely failed to use the character from Image 2 as requested
• − Only slightly modified the face of the original woman from Image 1 instead of replacing her with the man from Image 2

GPT Image 1 Mini

• + Successfully extracted and applied the exact outfit from Image 2, including the specific plaid pattern and coat style.
• + Maintains the skin vitiligo details on the hands and face.
• + The composition and perspective feel cohesive with the new body pose.

• − Changed the person's face and head shape significantly from Image 1.
• − The sand on the face and the specific hair patch from Image 1 were lost in the generation.
• − The transition between the neck and the clothing is slightly blurry.

Grok Imagine Image Pro

• + Perfectly preserved the original face and hair from Image 1.
• + High-quality rendering of fabric textures and lighting.
• + Maintains the exact wooden structure and background from the source image.

• − Completely ignored the instruction to use the outfit from Image 2, instead generating a generic royal costume.
• − Created white-skinned hands that do not match the person in Image 1.

GPT Image 1 Mini

• + Excellent close-up detail on the capybara's fur and expression.
• + Strong atmospheric lighting and authentic bokeh effect.

• − The passenger is very blurry and lacks detail.
• − Only one paw is clearly visible on the steering wheel.

Grok Imagine Image Pro

• + Perfect adherence to all prompt details including the jacket, two paws on the wheel, and the woman's expression.
• + Impressive text rendering on the hat with specific NYC medallion details.
• + Very clear composition that shows both subjects and the city environment equally well.

• − The capybara's paws look more like crab claws/talons than actual capybara paws.
• − The perspective from the dashboard looking back is slightly less intimate than Model A.

GPT Image 1 Mini

• + Excellent fur texture and lighting integration.
• + Natural, dynamic poses that feel like a real action snapshot.
• + Superior rendering of 'god rays' and soft, atmospheric depth.

• − The fox's anatomy is slightly wonky, particularly the mouth/teeth area.
• − The butterfly above the kitten lacks a shadow or realistic lighting interaction.

Grok Imagine Image Pro

• + Accurately included more varied butterflies and wildflowers.
• + Great interaction between animals, with the fox tumbling as requested.
• + Very sharp focus across all subjects.

• − The lighting feels more synthetic and 'over-processed' compared to Model A.
• − Added an extra kitten not explicitly requested, crowding the composition.
• − The fur texture looks slightly more like a digital painting than a photograph.