New ChatGPT image model: GPT Image 1.5 (4× faster)

If you’ve ever told an AI, “Just change the background,” and it replied with, “Sure! I also changed the face, the lighting, and your entire vibe 😄,” then you’re going to love what OpenAI just shipped.

On December 16, 2025, OpenAI rolled out a new, improved ChatGPT Images experience powered by its new flagship image model — and the big promise is simple: it makes the edits you asked for… and leaves the rest alone. 🙌🖼️

🚀 What’s new in the ChatGPT image model?

1) 🎯 Precise edits that preserve the important stuff

The update is built around a very satisfying idea:
Change only what you asked for, while keeping things like lighting, composition, and people’s appearance (facial likeness) consistent across edits. 🧠✅

So instead of “new jacket = new person,” you get:
“new jacket = same person, same scene, same vibe.” 😌🧥

2) ⚡ Up to 4× faster image generation

You can iterate quicker (aka: generate → tweak → regenerate without losing your patience 😅). OpenAI says images can generate up to 4× faster now.

Bonus: you can keep generating new images while others are still in progress — which is basically “multitasking, but for creativity.” 🏃‍♂️💨

3) 🧾 Better text inside images (finally!)

Posters, menus, mockups, “YouTube thumbnail text that doesn’t look cursed”… it’s improving. The model is described as being able to handle denser and smaller text better than before. ✍️📌

4) 🧩 A whole new “Images” space in ChatGPT (with shortcuts!)

Instead of writing the perfect prompt from scratch every time, there’s now a dedicated Images area in the ChatGPT sidebar to make exploring faster and easier. It includes:

Preset filters 🎛️
Trending prompts 🔥
And a one-time likeness upload so you can reuse your look without digging through your camera roll again 😄📸

OpenAI says the experience is rolling out for most users, with Business and Enterprise access coming later.

👩‍💻 For creators + devs: it’s in the API too

The same improvements are available via the API as GPT Image 1.5. OpenAI also notes it’s stronger at image preservation/editing than GPT Image 1 and better at keeping logos and key visuals consistent — which matters a lot for brand work and ecommerce catalogs. 🛍️🏷️

OpenAI says image inputs/outputs are 20% cheaper in GPT Image 1.5 vs GPT Image 1. 💸

🧠 How to get great results (copy-paste prompt tricks) ✨

✅ The “surgical edit” prompt (best for real photos)

Prompt:
🛠️ “Edit this image: [exact change]. Keep lighting, camera angle, background, and facial likeness the same. Do not change anything else.”

Examples:

“Remove the person in the background on the left.” 🧹
“Change the shirt to a black hoodie, keep everything else identical.” 🧥
“Make this look like a 2000s film camera photo, keep faces the same.” 🎞️

Another Example:

Prompt:

Make an old school golden age hollywood movie poster of a movie called 'codex' from the image of these two men. feel free to change their costumes to fit the times Change the names of the actors to Wojciech Zaremba (left) and Greg Brockman (right) Directed by Sam Altman, produced by Fidji Simo. A Feel the AGI Pictures Production.

Output:

🎬 The vibe shift: from “AI toy” to “useful creative tool”

A lot of the coverage describes this as OpenAI moving from novelty generations to something more like a practical creative studio — faster, more consistent edits, more usable outputs. 🧰🎨