Meta launched Muse Ember, its smallest and most efficient model, designed for repeated tasks such as tool calls, coding subtasks, routing, extraction, and long document parsing, starting at $0.10 per million input tokens and sitting on the efficiency frontier of the Artificial Analysis leaderboard.

1 min read

Meta Launches Muse Ember: Its Smallest, Cheapest Model for Repeated Agent Tasks

What's new?

FAQ

What is Muse Ember?

Muse Ember is Meta's smallest and most efficient model, designed for repeated tasks such as tool calls, coding subtasks, routing, extraction, and long document parsing.

How does Muse Ember's pricing compare to GPT-6 Luna and Claude Haiku 5.5?

For shorter prompts, all three models cost $0.10 per million input tokens and $0.50 per million output tokens. After a threshold, prices rise: Muse Ember above 200k tokens to $0.30 input and $1.50 output, GPT-6 Luna above 272k tokens to $0.20 and $0.75, and Claude Haiku 5.5 above 100k tokens to $0.50 and $2.50.

Should MENA AI teams adopt Muse Ember now?

It may be an attractive option for teams building AI agents or applications that rely heavily on repeated tasks, given its low cost and position on the efficiency frontier. However, the decision should depend on comparing actual performance in specific use cases, not price alone.

Source: Simon Willison (LLM & tools)

AI-assisted summary of the source above. High-confidence stories publish automatically; others wait for review. Facts that are not in the source are not added.