Meta Superintelligence Lab released Muse Ember on Monday, which the lab describes as its lightest and most cost-efficient AI model yet. In its announcement, Meta said Ember was "designed for complex reasoning execution and the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long document parsing."
Meta emphasized cost per task. It used Artificial Analysis's composite benchmark score to position Ember as cheaper and smarter than Anthropic's Haiku and OpenAI's Luna. The article says Ember costs 10 cents per million input tokens, which it describes as significantly cheaper than Haiku and Luna. It also says Ember is cost-competitive with the open-source models DeepSeek V4.1 Flash and GLM 5.3 Flash.
Ember is built to pair with a larger model, Muse Wonder, in an agentic setup. The article gives a customer-support example: a heavier model sets rules for a type of issue, and a lighter model applies those rules across tens of thousands of cases. Meta put it this way: "Muse Wonder plans, Muse Ember handles the volume."
Meta also trained Ember to handle text, images, and audio. According