X · X// 7h agoMODEL RELEASE

ChatGPT Images 2.0 flexes manga chops

Sam Altman showcased ChatGPT Images 2.0 with a manga-style image of himself and Gabriel Goh hunting for more GPUs. The broader launch positions OpenAI’s new image model as a more capable visual workflow tool, with thinking, web search, multi-image generation, better typography, and API access as gpt-image-2.

// ANALYSIS

This is a meme post, but the underlying release is serious: OpenAI is pushing image generation from novelty output toward usable design and storytelling infrastructure.

–Manga and comic examples matter because consistent characters, panels, readable text, and narrative continuity are where older image models often broke down
–“Thinking” before generation could make image models more useful for developers building asset pipelines, UI mockups, visual explainers, and automated creative workflows
–The GPU joke is doing real work: better image generation is compute-hungry, and social demand can turn a model launch into an infrastructure stress test overnight
–API availability as gpt-image-2 makes this relevant beyond ChatGPT users, especially for apps that need generated images with layout, text, or brand constraints

// TAGS

chatgpt-images-2-0openaiimage-genmultimodalapigpu

DISCOVERED

7h ago

2026-04-22

PUBLISHED

20h ago

2026-04-21

RELEVANCE

9/ 10

AUTHOR

sama