Mini toolsMuse Image
Muse Image is Meta Superintelligence Labs' agentic image generation model. It operates as an agent rather than a direct prompt to image mapper, invoking coding and web search tools to ground outputs in factual, real time information and to render accurate plots, QR codes, and figures. It exhibits emergent self refinement, locally editing drafts or regenerating from scratch when needed, and its quality scales with test time reasoning compute. Muse Image follows editing instructions precisely and composes elements from multiple reference images including people, objects, clothing, styles, and environments, interleaving text and images in prompts. It integrates with Muse Spark for joint agentic media generation and includes Content Seal invisible watermarking for provenance. It ranks No. 2 on Arena for text to image and image editing as of July 2026. Available in the Meta AI app, meta.ai, Instagram Stories, and WhatsApp.
New
Image
Overview
Muse Image is Meta's agentic image generation model. Rather than mapping prompts directly to pixels, it invokes search and coding tools, self-refines its own drafts, and scales reasoning at inference time. It follows instructions precisely, edits images with fine control, and composes new scenes from multiple reference images.
About Meta Platforms
We're connecting people to what they care about, powering new, meaningful experiences, and advancing the state-of-the-art through open research and accessible tooling.
Industry:
Technology, Information and Media
Company Size:
78865
Location:
Menlo Park, California, US
Website:
ai.meta.com
Last updated: July 8, 2026
