Key Info

Meta announced Muse Ember, its smallest and most efficient model, designed for complex reasoning execution and everyday agent work, available today on the Meta Model API and in Muse Code.

Highlights

  • Supports agent workflows: carrying tasks across apps, calling tools in sequence, running tests and making targeted code fixes
  • 1M-token context window with native understanding of text, images, audio and video
  • On the Artificial Analysis Intelligence Index, performs comparably to Claude Haiku 5.5 at roughly one-third the cost per task, and scores higher than GPT-6 Luna across all of Luna's reasoning settings
  • Priced at $0.10 per million input tokens and $0.50 per million output tokens