Zappa and Digitalmodelmanagement are not using it yet.
Announced today. A new image generator within ChatGPT. Improved text rendering, multilingual, sharper image editing. Previously, OpenAI had already released 4o-image generation as standard in ChatGPT. Now, there's a significant step forward.
Interesting. Not because everything changes. But because within ChatGPT itself, you can now work much more precisely with images.
What Is Now Possible
Generate an image from a precise prompt
Use an existing image as visual input or reference
Control style, composition, and details much more sharply
Make targeted adjustments instead of starting over
Especially the latter is crucial. Reference image plus tight prompt structure.
Not: "create a beautiful fashion image."
But:
Where It Becomes Interesting for Us
Our sequence remains the same: first identity, then aesthetics. Never the other way around.
This type of prompt works directly with that, see the start of the prompt at the bottom of this blog.
You don't use ChatGPT Image 2.0 as a "just make something beautiful" machine. You direct it precisely. That's where it becomes useful.
What This Does Not Mean
We are not changing our methodology.
What remains:
That is our foundation.
Why It Is Still Interesting
Image 2.0 demonstrates that you can seriously test within ChatGPT with:
identity lock
portrait quality
skin, tension, and facial detail
image construction from language plus reference
An extra layer in the process. Not a replacement for our stack.
🧠 quite good at building identity
🎯 understands skin, tension, face — high-fashion level (full testing to follow)
❌ no memory, no consistency engine
Conclusion
For us, Image 2.0 is now primarily:
GPT launch tonight, 21 April, 21:00 NL time: https://www.youtube.com/watch?v=sWkGomJ3TLI
See small differences, first image created by: Higgsfield / Nano Banana Pro
High-fashion studio editorial photograph, tight head-and-shoulders portrait, full color.
Torso angled slightly toward camera-right, shoulders staggered with the right shoulder positioned marginally forward.
Head turned toward camera with slight downward tilt, chin subtly lowered, gaze directed straight into lens.
Right shoulder covered by structured blazer lapel; left shoulder partially visible beneath blazer edge.
No visible arm elevation within frame; blazer fabric resting naturally along shoulder line without lift.
Visible contact points: layered garments resting against collarbone and shoulder.
Tailored blazer with structured lapel and visible weave texture; light blue collared shirt beneath with open neckline and visible button placket; fabric surfaces show natural grain, seam definition, and fold formation; skin rendered with natural tonal variation and visible pore-level detail.
Hair remains identity-locked to the reference image and is described only through physically verifiable motion and interaction. No colour, length, or style descriptors are permitted.
Strong directional warm light entering from camera-left creating pronounced light streaking across the facial plane and partial optical flare; light rakes across forehead, cheekbone, and nose bridge, leaving the opposite side in controlled soft shadow; high contrast with feathered shadow edges; highlight roll-off visible along cheek and lower lip; subtle non-linear highlight bloom interacting with hair edge and blazer shoulder; warm skin tones contrasted against a neutral beige background; high-dynamic range fall-off in brightest streak zones without clipping.
Minimal studio environment with smooth beige backdrop; shallow depth separation; mild telephoto compression; photographic layering with intentional optical flare passing across mid-frame.
Pure studio photography. Editorial realism. No cinematic effects.
And from now on: this result from ChatGPT Images 2.0
High-fashion studio editorial photograph, tight head-and-shoulders portrait, full color.
Torso angled slightly toward camera-right, shoulders staggered with the right shoulder positioned marginally forward.
Head turned toward camera with slight downward tilt, chin subtly lowered, gaze directed straight into lens.
Right shoulder covered by structured blazer lapel; left shoulder partially visible beneath blazer edge.
No visible arm elevation within frame; blazer fabric resting naturally along shoulder line without lift.
Visible contact points: layered garments resting against collarbone and shoulder.
Tailored blazer with structured lapel and visible weave texture; light blue collared shirt beneath with open neckline and visible button placket; fabric surfaces show natural grain, seam definition, and fold formation; skin rendered with natural tonal variation and visible pore-level detail.
Hair remains identity-locked to the reference image and is described only through physically verifiable motion and interaction. No colour, length, or style descriptors are permitted.
Strong directional warm light entering from camera-left creating pronounced light streaking across the facial plane and partial optical flare; light rakes across forehead, cheekbone, and nose bridge, leaving the opposite side in controlled soft shadow; high contrast with feathered shadow edges; highlight roll-off visible along cheek and lower lip; subtle non-linear highlight bloom interacting with hair edge and blazer shoulder; warm skin tones contrasted against a neutral beige background; high-dynamic range fall-off in brightest streak zones without clipping.
Minimal studio environment with smooth beige backdrop; shallow depth separation; mild telephoto compression; photographic layering with intentional optical flare passing across mid-frame.
Pure studio photography. Editorial realism. No cinematic effects.
For ZAPPA the core remains unchanged.
But next week we will test and update everything in Image 2. More to follow.
Peter
// IDENTITY LOCK (NON-NEGOTIABLE)
Use the provided reference image as the exact identity source.
Face structure, proportions, eyes, nose, mouth, bone structure and skin characteristics must remain unchanged.
No beautification.
No reinterpretation.
No deviation.
No blending with generic faces.
Hair remains identity-locked and is only affected by physical interaction (light, gravity, movement).
No descriptive override allowed.
---
SCENE / COMPOSITION
[ your existing prompt here ]
// Paste your code here...
