Meta releases Muse Ember for repeated agent and coding work
Original title: Our Smallest, Most Efficient Model: Muse Ember
It is on the Meta Model API at $0.10 per million input tokens and $0.50 per million output tokens.
Meta introduced Muse Ember as its smallest, most efficient model for tool calls, coding subtasks, routing, extraction, and long-document parsing. Meta says it delivers more intelligence per dollar than Claude Haiku 5.5 or GPT-6 Luna and cites Artificial Analysis scores, including AutomationBench-AA at 37.8 for Ember at max reasoning. It has a 1M-token context window and up to 512K output tokens. Several cost figures are estimates at proposed pricing.
Unconfirmed: Efficiency rankings and several scores are Meta's claims; some Ember costs are estimates.
- List pricing is $0.10 per million input and $0.50 per million output tokens.
- Context is 1M tokens, with up to 512K output tokens.
- Meta reports Terminal-Bench 4.0 at 40.9 for Ember at max reasoning.
- Box says early tests included contract scoring and candidate screening.
- Vendors
- Meta
- Products
- Meta Model API
- Organizations
- Artificial AnalysisBox
- AI models
- Muse EmberClaude Haiku 5.5GPT-6 Luna
From the original article
7 min read at research.meta.ai
Today, we're introducing Muse Ember, our smallest and most efficient model designed for complex reasoning execution and the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long document parsing. Muse Ember delivers more intelligence per dollar than Claude Haiku 5.
By dvaplima · research.meta.ai
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.