Meta launches Muse Ember, its smallest and cheapest agent model

Meta’s smallest Muse model is built for the repetitive work agents do all day, at $0.10 per million input tokens.

In this brief: 3 sections 1 min read

- • Built for tool calls, coding subtasks, routing, extraction, and long-document parsing
- • First model trained using Muse Wonder Full, Meta’s largest model
- • Available now as muse-ember-1.0 on the Meta Model API and in Muse Code
- • $0.10 per million input tokens, $0.004 cached, $0.50 per million output tokens
- • 1M-token context window, up to 512K output tokens in one call
- • Headline price applies up to 200K input tokens; larger requests cost more
- • Fourth model in the Muse family, after Spark (April), Glimmer (August), and Wonder (October)
- • Meta says it beats Claude Haiku 5.5 and GPT-6 Luna on intelligence per dollar
- • Scores level with Claude Haiku 5.5 on Artificial Analysis’s Intelligence Index

Useful?