Meta enters AI image model race in bid to court advertisers and subscribers

In a bold move to enhance its offerings for creators and advertisers, Meta has introduced Muse Image, an innovative artificial intelligence model designed for image creation. This announcement comes as part of the company's ongoing efforts to attract more users and expand its advertising capabilities. Previously known by its codename Mango, Muse Image represents the second significant launch from Meta's Superintelligence Labs, led by Alexandr Wang, who was also behind the April release of the Muse Spark large language model, which followed the Llama model family. Users will have free access to Muse Image through the Meta AI app and website, as well as via direct messages on WhatsApp and Instagram Stories. However, for power users and content creators aiming to generate numerous AI-crafted images or utilize advanced features, subscription plans introduced in May will be necessary. Users who exceed their free usage limits can opt for a Meta One subscription or wait for their limits to reset, according to company guidelines. The new AI model is also set to enhance Meta's advertising capabilities, particularly through its AI-powered Advantage Plus service, which aims to simplify the process for brands looking to develop creative marketing content. Meta has collaborated with various businesses to ensure Muse Image meets the needs of advertisers. As per a recent blog post, the model integrates native reasoning into the creative process, allowing for element adjustments, style swaps, and variations tailored to advertisers' needs, thereby creating high-quality ad versions with fewer revisions. As Meta ventures into this competitive space, they acknowledge that rivals OpenAI and Alphabet have already established a foothold with their own image-generation technologies. Google's Nano Banana, for instance, gained significant popularity upon its release last fall. Internal benchmarking tests indicate that while Muse Image does not yet surpass OpenAI's latest GPT Image 2 model, it does outperform the Nano Banana 2 in various editing tasks. Historically, Meta has relied on third-party AI models like Midjourney and Black Forest Labs for its image and video generation features. With the debut of Muse Image, the company aims to lessen its dependence on these external technologies. Additionally, Meta is gearing up to launch an AI video generation model named Muse Video, which promises competitive capabilities in terms of prompt adherence, visual quality, and consistency over time. Muse Image is expected to roll out across Facebook and Messenger, along with expanded access within Instagram and WhatsApp later this year.