Meta's Muse Ember Is Its Smallest, Cheapest Model, Built For AI Agents
0:00 / 0:41
News
Meta's Muse Ember Is Its Smallest, Cheapest Model, Built For AI Agents
calendar_today Date:
schedule Duration: 0:41
database
Summary Report
Meta releases Muse Ember, its smallest Muse model, built for high-volume agent work at $0.10 per million input tokens.
- 01. Built for agent subtasks; 1M-token context; text, images, audio and video
- 02. Beats Haiku 5.5 and GPT-6 Luna on most of Meta's tests; $0.10/$0.50 per million tokens
Meta has released Muse Ember, the smallest model in its Muse family, built for the repetitive work agents do all day: tool calls, coding subtasks, routing, extraction and long document parsing. It has a 1M-token context window and understands text, images, audio and video. On Meta's own charts it beats Claude Haiku 5.5 and GPT-6 Luna on most agent and coding tests, though Haiku wins on FrontierCode. It costs $0.10 per million input tokens and $0.50 output, and is available now on the Meta Model API and in Muse Code. Meta hasn't said anything about open weights.
Meta Data
Company:
LLM:
Model:
