Muse Ember

Meta's smallest and cheapest Muse model, built for high-volume agent work such as coding subtasks and tool calls.

Takes text, image, audio, and video as input.
Available in Muse Code, included in all Muse Code plans, and on the Meta Model API as muse-ember-1.0 through an OpenAI-compatible endpoint at api.meta.ai/v1.
Pricing per 1M tokens is $0.10 input and $0.50 output up to 200k input tokens, and $0.30 input and $1.50 output above that.
Meta reports a score of 36 on the Artificial Analysis Intelligence Index and says it beats similar-size models on AutomationBench-AA and Terminal-Bench 4.0.
Meta's guidance is to pair it with larger Muse models for deeper planning and reasoning. The page does not state a context window or rate limits.

Repeated, bounded agent work at scale, like coding subtasks, tool calls, routing, and document analysis.

Use it for the cheap, fast steps and hand deeper planning to Muse Wonder or Muse Spark.

Developers building agents who need to keep cost per task low.

Join the newsletter to get ahead with AI.

I share strategies, new modules, free guides, and tools to get you better at using AI in your work.