Meta Releases Muse Ember, a Low-Cost Model for Agent and Coding Tasks

At a glance

Meta has released Muse Ember, its smallest AI model. It is designed for tasks that AI agents repeatedly perform, such as tool calls, coding subtasks, and document processing. It is priced at $0.10 per million input tokens.

Key points

1. 1

   Pricing is based on tokens (units that AI uses to split up text), so costs rise as usage grows.

2. 2

   In Meta's evaluations, Ember costs about one-third as much per task yet scored close to the top score of Claude Haiku 5.5.

3. 3

   Ember can handle text, images, audio, and video together, so it can work from screenshots or meeting recordings.

4. 4

   However, Haiku led on FrontierCode, a difficult engineering evaluation, so Ember cannot be considered better at all coding tasks.

Why it matters

The more AI agents are used, the higher the costs become, so cheaper and faster models can help companies and developers reduce their burden.

Also reported by

• TechCrunchMeta releases Muse Ember, a new lightweight model for agent execution
• The DecoderMeta launches Muse Ember, a small AI model built for the repetitive work agents actually do