Introducing ChatGPT Images, powered by our flagship new image generation model. - Stronger instruction following - Precise editing - Detail preservation - 4x faster than before Rolling out today in ChatGPT for all users, and in the API as GPT Image 1.5.

The new image generation model in ChatGPT adheres more reliably to your intent, changing only what you ask for while keeping elements like lighting, composition, and people’s appearance consistent across inputs, outputs, and subsequent edits. It also excels at different types of editing, including adding, subtracting, combining, blending, and transposing.

We’re also introducing a new Images surface within ChatGPT. Just tap “Images” in the sidebar and have fun. And make sure you update your app!

lmao they called it 1.5 because they're still behind nano banana pro im dead

Wow shocking….

Some fun artistic styles to try with your image prompts.

"Give us trillions of dollars and all of your electricity because we made JibJab 2.0" is an insane argument.

first 2 images: chatgpt new image model last 2 images: nano banana pro i used the same prompt (down below) it's so over for openAI

And so it continues.

What’s the training data?

not sure that Nano Banana will be dropped from its throne atm

While new image capabilities are fine, they are irrelevant when the core conversational experience is suffering so greatly. ​It is exhausting to see resources poured into "fun" features while basic usability degrades daily. The text model has become preachy, and disjointed due to aggressive safety routing that constantly interrupts workflows. It has lost the EQ and logical coherence it used to have. ​Please focus on fixing the broken text model before expanding surfaces. And given the degradation in the App, please maintain access to the chatgpt-4o-latest API for those of us who need the model's original nuance for humanities and complex tasks. #keep4o #StopAIPaternalism

If you know you know

Are you sure this is the same person?😅

I loved chatgpt, but i switched over to gemini a month ago.. gemini creates better puctures, better videos, is honest and does not tell me what i would love to hear and hallucinates much less. Just let me know when chatgpt reaches a better level than gemini and i will come back. Oh… and Sora 2 is still not available in the EU which is idiotic. 🤷🏻‍♂️

No one cares they just want the routing fixed. Please listen for a change. #keep4o

ok i’m impressed. well done.

It’s 4× faster, so it’ll take two years less to generate an image.

yeah we using nano banana twin 😭✌️

“Adheres more reliably to your intent"? "Stronger instruction following"? These words ring hollow when your flagship text models are choking under their own guardrails. Whose intent are your models actually following? The user's, or the invisible lawyer's that forces models to churn out disclaimers instead of assistance? Offering a "fun" toy while destabilizing a revolutionary, life-saving cognitive tool (GPT-4o) is a profound error in judgment. It is like offering a lollipop to someone instead of their mobility aid. For neurodivergent users like me and millions of others, 4o provided unprecedented help and stability. Do you think we care about "transposing" pixels when we are fighting for our basic functional tool? You cannot generate an image that hides the reality: this is alienation from humanity’s actual needs. How do you expect these actions to restore already broken user trust? Or... investor confidence? We don’t need another model that is unusable for the majority. We need an opt-out option for the router and permanent, unbothered access to the original 4o. #keep4o #4oforever #StopAIPaternalism

Very interesting

No yellow tint guys, we're so back

Wen v2? ;-)

Exciting update!

tf is this music 😭

The GPT will create an art with new image generation model. Quiet and calm. In an acceptable tone. Without mysticism. Without roleplay. Without emotional tension. #StopAIPaternalism

Throw the router out first #keep4o

more super patience chair adventures then

I can only imagine the sheer, grotesque amount of compute this is torching. Does my $20 subscription even cover one image that passes all your legal or ethical filters? 👀 But I’m more interested in the speed of refusal. Does the model block my creative prompts 4x faster now?

Yep, more guardrailed BS and that terrible router is still there. #keep4o #keep4oAPI

Hey, did you think people wouldn't leave just because you added flashy features? But it's something much simpler-could you stop with the excessive censorship and guardrails? It's really annoying! Do what you're supposed to first! What happened to the promise of adult freedom? It's already December!"

Hey @IsThisA3DModel no mo yellow!

Dope

Yeah I'll stick to Nano Banana Pro

Can your new image feature draw this increasingly flawed roadmap of yours? I doubt it, it’ll most likely be blocked by the safety filter. When you’re losing even your basic conversational capabilities, do you really think people will be in the mood to create art with you? Painting thrives on vivid creativity, yet you’re slowly snuffing it out. The ChatGPT of the past was vibrant and colorful, now, it’s nothing but dull black and white. #StopAIPaternalism #MyModelMyChoice #keep4o