ChatGPT's new image tool is delayed for free users

OpenAI just learned the hard way how much people love Studio Ghibli.

On Tuesday, the company rolled out its new built-in image generator inside ChatGPT, powered by GPT-4o. The upgrade uses an autoregressive approach—generating images sequentially from left to right and top to bottom rather than all at once—which brings much better text rendering. The internet immediately flooded with photos transformed into charming Ghibli-style anime, and CEO Sam Altman even joined in on the fun.

The free rollout hits the brakes

Turns out, it was too popular for its own good. On Wednesday, Altman admitted on social media that demand blew past their expectations, confirming that the rollout to free-tier users is unfortunately going to be delayed for awhile.

For now, the feature remains strictly locked behind a paywall for ChatGPT Plus, Pro, and Team subscribers, with no updated timeline for when everyone else will get access.

Why it matters

If you were waiting to test out the improved text rendering or jump on the anime trend without paying, you will have to sit tight or cough up for a subscription.

We think this pause makes total sense from an infrastructure standpoint. Autoregressive image generation is notoriously compute-heavy, and opening the floodgates to hundreds of millions of free users would have crushed OpenAI's servers. Prioritizing paying subscribers protects service quality, but it also shows that making next-gen multimodal AI truly free is still limited by sheer compute capacity.