On Tuesday, Meta Platforms, Inc. introduced Muse Image, a new artificial intelligence model designed for image creation, as the company aims to attract creators and advertisers.
This AI technology, internally codenamed Mango, is the second major product from the Meta Superintelligence Labs, led by Alexandr Wang, who previously oversaw the Muse Spark large language model launched in April, which succeeded the company's earlier Llama series.
Muse Image will be made available to consumers for free via the Meta AI app and website, WhatsApp direct messages, and Instagram Stories. Advanced users and creators will need to subscribe to one of Meta's monthly subscription plans launched in May to generate a high volume of AI images and access specific features. The company stated that users who reach the free limit can purchase a Meta One subscription or wait for the limit to reset.
Muse Image will also support a dedicated image generation tool for advertisers as part of the Meta AI-powered Advantage Plus service, helping brands more easily develop ad creative for marketing campaigns and automate certain tasks. Meta indicated it has been collaborating with businesses and advertisers during the development of Muse Image.
In a corporate blog post, Meta stated: "Muse Image integrates native reasoning capabilities into the creative process, enabling the adjustment of elements, switching of styles, and generation of variants based on an advertiser's concept, thereby producing high-quality, brand-aligned ad variations with fewer iterations."
The new image generation model and its monetization efforts signal that Meta is attempting to expand beyond its core online advertising business to open up new revenue streams, offsetting its massive investments in AI-related infrastructure.
Alphabet and OpenAI are ahead of Meta in offering similar image generation models, with Google's Nano Banana gaining significant consumer popularity after its release last autumn.
Meta also disclosed internal benchmark results showing Muse Image lags behind OpenAI's latest GPT Image 2 model on tasks like editing single and multiple images but performs better than the Nano Banana 2 model.
The social media giant had previously used third-party AI models, such as Midjourney and Black Forest Labs, to power various image and video generation features within its Meta AI app and website. The company stated it plans to leverage its new AI model to reduce reliance on similar third-party technologies.
Meta also plans to release an AI video generation model called Muse Video in the future, adding in a technical blog that the model "demonstrates competitive performance in prompt adherence, visual fidelity, and temporal consistency."
Muse Image is scheduled to launch on Facebook and Messenger later this year and will expand to more areas within Instagram and WhatsApp services.
Comments