Meta has officially unveiled Muse Image, its newest AI image generation model developed by Meta Superintelligence Labs (MSL). It is the first image-generation model created by the new AI division and is set to power visual AI features across Meta’s ecosystem. Muse Image goes beyond basic text-to-image generation. The model can accurately follow complex prompts, edit existing images, combine multiple reference images into a single composition, and modify visuals using sketches or annotations.
According to Meta, Muse Image uses an agentic approach to image generation. Instead of directly converting prompts into images, it can leverage search and coding tools to improve accuracy. For example, the model is capable of writing and executing code to generate charts, diagrams, and QR codes before incorporating them into the final image. Muse Image also works alongside Muse Spark, Meta’s flagship AI model, enabling the creation of animated GIFs, interactive visuals, and image-rich web experiences. Alongside the launch, Meta also previewed Muse Video, its upcoming AI video generation model built on the same foundation.
Muse Image is now available through the Meta AI app and meta.ai. It is rolling out to Instagram Stories in the United States and to WhatsApp in selected countries, with support for Facebook and additional Meta services planned in the future.
