This is an AI image generation comparison for
text-to-image
prompt:
An expressive fauvist oil portrait of a French sommelier , bust portrait, simplified anatomy, bold unnatural colour planes across the face (green, yellow, violet shadows), thick impasto oil paint, visible brushstrokes, raw pigment texture, asymmetrical colour modelling, naive early modernist portrait, decorative colour-block background, flattened perspective, 1905 avant-garde painting, museum oil painting, textured canvas, painterly imperfections...
Log in to see full prompt.
Tested: March 13, 2026
Prompt by glasskite.bsky.social (see link to original)
Tested: March 13, 2026
Not bad
Tested: March 13, 2026
Banana 2 doesn't disappoint
Tested: March 13, 2026
Reve fast
Tested: March 13, 2026
Done well with this prompt
Tested: March 13, 2026
PRO version.
Tested: March 13, 2026
Frankly, 4.5 seems superior to 5 Lite for this prompr
Prompt exploring whether AI video or image models can convincingly recreate a Fauvist 1905-style oil portrait—bold unnatural facial colours, heavy impasto brushstrokes, and naive early modernist anatomy in a museum-style painted bust of a French sommelier. A strong test of painterly texture and art-historical style fidelity.
Is the subject clearly a French sommelier shown as a bust portrait (head and upper torso only)?
Are the facial colours unnatural and fauvist, with distinct planes such as green, yellow, or violet shadows?
Is the anatomy simplified or slightly naive rather than realistically rendered?
Are thick impasto oil paint textures clearly visible on the surface?
Are individual brushstrokes and raw pigment texture visibly present across the face and clothing?
Does the colour modelling appear asymmetrical, with bold mismatched tones on different sides of the face?
Is the background composed of decorative colour blocks rather than a realistic environment?
Is the perspective flattened, giving the painting a stylized early-modernist look?
Does the canvas show painterly imperfections rather than smooth digital gradients?
Is the overall style consistent with a 1905 Fauvist / early avant-garde museum oil painting?
Check out the results from Flova.ai (Midjourney 7) vs Pollo AI (Pollo Image 2.0) vs GROK (Grok Imagine v0.9 Image) vs Freepik (Seedream 4.5) vs Freepik (Nano Banana 2) vs Freepik (Reve Image 1.5) vs Freepik (Imagen 4 Ultra) vs Recraft (Recraft V4) vs Freepik (Seedream 5.0 Lite) for similar or identical prompts side-by-side.