How Did the Trend Begin?
On Tuesday, OpenAI introduced a groundbreaking feature within GPT-4o, captivating users with its ability to replicate the iconic Studio Ghibli aesthetic. While the model was designed for high-accuracy and stylized visuals, its uncanny ability to mimic Hayao Miyazaki’s legendary art style quickly turned it into a viral sensation.
Within hours of its release, users began experimenting with the image generator, testing its capabilities with various prompts. It soon became apparent that the AI excelled at creating Ghibli-style artwork. Social media platforms like X, Instagram, and Reddit were soon flooded with AI-generated Ghibli renditions of selfies, pets, fictional characters, and even everyday objects like coffee cups and bicycles.
The trend spread rapidly, with users sharing their creations under hashtags like #GhibliStyle and #AIGhibli. Some even reimagined iconic pop culture moments—transforming celebrities, historical figures, and internet memes into Ghibli-style characters.
It's been 24 hours since OpenAI unexpectedly shook the AI image world with 4o image generation.
— Barsee 🐶 (@heyBarsee) March 26, 2025
Here are the 14 most mindblowing examples so far (100% AI-generated):
1. Studio ghibli style memespic.twitter.com/E38mBnPnQh
Even OpenAI CEO Sam Altman joined in. On Wednesday, he humorously posted on X:
“Be me. Grind for a decade trying to help make superintelligence to cure cancer or whatever, mostly no one cares for the first 7.5 years, and then for 2.5 years everyone hates you for everything. Wake up one day to hundreds of messages: ‘look I made you into a twink Ghibli style haha.’”
>be me
— Sam Altman (@sama) March 26, 2025
>grind for a decade trying to help make superintelligence to cure cancer or whatever
>mostly no one cares for first 7.5 years, then for 2.5 years everyone hates you for everything
>wake up one day to hundreds of messages: "look i made you into a twink ghibli style haha"
Altman even changed his profile picture to an AI-generated Ghibli-style version of himself.
ChatGPT-4o’s Image Generation
OpenAI’s latest update, powered by GPT-4o, introduces in-chat image generation through a feature called “Images in ChatGPT.” While available to all users, free-tier users face some limitations due to high demand.
However, the rollout hasn’t been without hiccups. OpenAI initially delayed access for free users and had to address a bug where the AI would generate images of “sexy men” but not “sexy women.”
Despite these early challenges, GPT-4o significantly improves image accuracy, text rendering, and object relationships. Unlike DALL-E, which generates images all at once, GPT-4o constructs them piece by piece—enhancing realism and detail.