If you’re a Google Bard user, you know that the popular Gemini-powered chatbot lacks the ability to generate images, like some similar products from rival AI-centric companies like OpenAI. However, the search giant eventually rolled out Bard’s ability to generate images for all users.

Now, when you open Google Bard by visiting bard.google.com, you’ll see a small prompt that lets you generate images. “Start the prompt with ‘Create an image of’ or ‘Generate an image of’ and describe the image you have in mind,” Google notes.

Once you have the image, you can click “Generate More” for more options. Once the image is generated, you can quickly click the download button to get the image offline.

We tried the tool and in our experience it works as advertised. Just submit a tip that includes key details like rendering style, characters, and even lighting.

In our tests, we asked Bard to generate realistic images of black holes, and the results were striking. We went a step further and asked if it could generate what’s inside a black hole, but the chatbot was smart enough to show that light can’t escape from a black hole; it’s impossible to know what it’s like inside.

In other words, Google Bard’s new features apply to the Imagen 2 text-to-image model and are not currently available to Gemini Ultra. Additionally, generating images with Bard is completely free, unlike ChatGPT Plus, which is only available as a paid subscription service. This is a significant advantage for Bard as more people can use its image generator and spread the word.