The new model replaces the previous Llama-based lineup and functions as an agentic system. Alexandr Wang, who leads the Superintelligence Labs, explained that the software pairs with the Muse Spark large language model to reason through user prompts and browse the web before rendering an output. Beyond simple image generation, the system allows users to redesign interior spaces using Facebook Marketplace listings and modify existing photos through direct on-screen drawing tools.
Meta Integrates Muse Image Model Across Social Platforms
Users on Instagram can now pull other people into AI-generated photos, as Meta rolls out its first image model developed by the Superintelligence Labs division. The Muse Image tool, which powers creative features across WhatsApp and Meta AI, uses public profile data to build visuals based on tagged usernames.
Meta is also preparing a Muse Video model to follow this launch, promising higher temporal consistency and prompt adherence. While the technology offers capabilities like creating custom postcards and invitations, the company has implemented controls for users to restrict how their content is repurposed by the AI. These features are debuting through 30 new effects for Instagram Stories, currently available in the United States before a planned international expansion across the Meta ecosystem.
Comments (0)
No comments yet. Be the first!