Meta Launches Muse Ember, a Low-Cost AI Model for Faster Agent Tasks

Meta has launched Muse Ember, its smallest and most cost-efficient artificial intelligence model, designed to handle repetitive tasks performed by AI agents. Introduced on November 9, 2026, the model focuses on activities such as tool calls, coding subtasks, data extraction, request routing and long-document processing. Meta says it is built to deliver reliable performance at a lower cost, particularly for businesses running large volumes of automated tasks.

The company says Muse Ember delivers strong performance compared with competing models, including Anthropic’s Claude Haiku 5.5 and OpenAI’s GPT-6 Luna, while costing less per task. Its pricing starts at $0.10 per million input tokens and $0.50 per million output tokens. Meta’s published benchmarks show strong results in agent operations and coding evaluations, although the model scored below Claude Haiku 5.5 on the FrontierCode 2.0 assessment.

Muse Ember also supports multiple forms of information, including text, images, audio and video. It features a context window of up to one million tokens and can generate up to 512,000 output tokens, allowing it to process large documents and entire codebases. Meta says the model can work alongside its more powerful Muse Wonder system, with Wonder handling complex planning while Ember carries out repeated tasks at scale.

The new model is available through Meta’s Model API and Muse Code, giving developers access to its capabilities for building automated workflows and AI-powered applications. Its release expands Meta’s Muse model family and adds to growing competition among technology companies to develop affordable AI systems. By focusing on speed, lower operating costs and high-volume execution, Meta is positioning Muse Ember as a practical option for businesses seeking to expand their use of AI agents.