Muse Image: Meta's New AI Model for Image Generation Lands on Instagram and WhatsApp
Meta has announced the launch of Muse Image, a new artificial intelligence model dedicated to image generation, aimed at expanding its offerings for content creators, advertisers, and platform users. The project represents the second major release developed by Meta Superintelligence Labs, the division led by Alexandr Wang, following the presentation of the Muse Spark language model in April.
Internally known during development by the code name "Mango", Muse Image will be available for free through the Meta AI app and website, as well as through direct messages on WhatsApp and Instagram Stories. Users wishing to create more AI-generated images or use additional features will need to subscribe to one of the new subscription plans introduced by Meta in May. Once the free limit is reached, users can either wait for quota restoration or upgrade to the Meta One subscription.
As highlighted by Meta, users can mention (@) other Instagram accounts in Muse Image prompts, allowing the AI model to incorporate those accounts' images into the results. Meta states that "tagging a username allows Meta AI to use public photos to create an image," but emphasizes that users can control how their content is reused by the AI.
The new model will also be used in Advantage Plus advertising tools, the AI platform dedicated to businesses. According to Meta, Muse Image will allow for modification of graphic elements, style changes, and the production of numerous creative variants starting from the original material provided by advertisers, reducing the number of revisions needed to achieve brand-consistent content.
Muse Image: The new Meta platform could soon present a version for videos. The company explained that it collaborated with businesses and advertising agencies during the model's development and expects to gradually roll out new features for advertisers in the coming weeks. With this initiative, Meta also aims to diversify its revenue sources beyond traditional online advertising, leveraging investments made in AI infrastructure.
In the sector, competitors like OpenAI and Alphabet have already introduced similar tools. Meta also released some results from its internal tests, where Muse Image was indicated as inferior to OpenAI's latest GPT Image 2 model but superior to Google's Nano Banana 2 model in some editing tasks involving both single and multiple images.
Thus far, Meta has relied on technologies developed by companies like Midjourney and Black Forest Labs for various image and video generation functions. With Muse Image, the goal is to decrease this dependence. The company also confirmed that Muse Video will also arrive in the future, a model dedicated to AI-driven video creation.