Meta Superintelligence Labs has launched Muse Image and previewed Muse Video, its first media generation models. Muse Image operates as an agentic model that uses coding and search tools, performs self-refinement during generation, and scales with test-time compute. It supports multi-reference image composition, iterative editing, and integrates with the broader Meta ecosystem including Instagram and WhatsApp. Muse Video, built on the same pretraining base, offers competitive text-to-video generation with native audio support. Both models include Content Seal, an invisible watermarking system for AI provenance. Muse Image currently ranks No. 2 on Arena for text-to-image and editing benchmarks, while Muse Video ranks No. 3 for text-to-video.
