Revealing the new feature in an X post, Musk said, “You can upload any image to Grok for analysis, from medical tests to video games! Just tap the + button or paste the image into the entry bar.”
Responding to the post, one user said how they used Grok in order to reveal a secret passage in the game Super Mario Bros.
“This is going to be an absolute game changer,” said another.
Grok, in collaboration with Black Forest Labs, could already generate images using the Flux 1 model, although free users can only ask up to 10 questions every two hours and are limited to three image analyses and four image generations per day, according to the chatbot’s guidelines.
Discover the stories of your interest
On the other hand, subscribers to X’s Premium and Premium+ plans, priced at $7 and $14 per month, respectively, enjoyed expanded access, with the ability to ask up to 50 questions daily.
Competitors like OpenAI, Meta and Google, among others, have their own image generators as well. OpenAI has its text-to-image tool called DALL·E 3, which uses the AI chatbot ChatGPT to generate images. Google has Gemini 2.0 that can generate images and audio in addition to text. Meta had rolled out the Meta AI personal assistant in June 2024, across all its platforms – WhatsApp, Facebook and Instagram – which offers text and image generation.
Last month, X had introduced a new image generator integrated into its Grok assistant named Aurora. It went live briefly and then vanished for some users shortly after. Similar to Flux, Aurora operated with minimal restrictions. It was able to generate images of public figures and copyrighted characters, such as Mickey Mouse, without issue. While it avoided creating explicit nudes during tests, it did permit graphic content.
Musk released Grok back in November 2023 in order to compete with AI assistants of other tech majors like OpenAI, Google and Meta.