When Google introduced its image generator for Gemini in early February, and the tool immediately started cranking out racially biased pictures of racially diverse Nazis and “founding fathers.”

“We’re aware that Gemini is offering inaccuracies in some historical image generation depictions,” a Google statement on X (formerly Twitter) from February 21 reads. “We’re working to improve these kinds of depictions immediately. Gemini’s AI image generation does generate a wide range of people. And that’s generally a good thing because people around the world use it. But it’s missing the mark here.”

It took the company a few months, but Gemini can once again generate images of humans. But only if you pay for it. To start, Google plans to offer this capability only to Advanced, Business, or Enterprise subscribers. The company did not disclose when it would be available to all users.

“Gemini Advanced gives our users priority access to our latest features,” a Google spokesperson told TechCrunch. “This helps us gather valuable feedback while delivering a highly anticipated feature first to our premium subscribers.”

The company has also announced that it is finally rolling out its Gems feature, which was first introduced at I/O in May. Gems are Gemini’s version of ChatGPT’s GPT, down to the truncated naming structure. They’re essentially mini-chatbots that perform only a few specialized functions, like giving gardening advice, cooking and nutrition tips, or trip planning. Since they’re smaller, they can infer responses more quickly than bringing the entire weight of the larger language model to bear.

Gemini Advanced, Business, and Enterprise subscribers will soon be able to create their own Gems on their mobile and desktop devices in over 150 countries and “most” languages.