Key facts
- Meta priced Muse Ember at 10 cents per million input tokens, below frontier competitors Haiku and Luna.
- Meta said Ember is its lightest and most cost-efficient model and is built for agent tasks.
- Meta designed Ember to pair with its larger Muse Wonder model in an agentic setup.
- Meta trained Ember to handle text, images and audio without a separate model.
- Ember is the fourth model under Meta's Muse brand, after Spark, Glimmer and Wonder.
Meta Superintelligence Lab released a new AI model on November 9, 2026 called Muse Ember, which the lab described as its lightest and most cost-efficient model yet. In its release, Meta said Ember was "designed for complex reasoning execution and the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long document parsing."
Meta emphasized Ember's cost per task, citing Artificial Analysis's composite benchmark score to present Ember as cheaper and smarter than competitors including Anthropic's Haiku and OpenAI's Luna. At 10 cents per million input tokens, Ember is cheaper than those frontier competitors and cost-competitive with open-source models DeepSeek V4.1 Flash and GLM 5.3 Flash, according to the report.
Meta designed Ember to pair with a larger model such as its Muse Wonder in an agentic harness. In a customer-support agent, for example, a heavier model could set rules for handling an issue and a lighter model apply them across tens of thousands of cases. Meta put it in the announcement: "Muse Wonder plans, Muse Ember handles the volume."
Meta also trained Ember to handle text, images and audio. The company said agents "can read screenshots, charts, and scanned documents, or work from voice memos, meeting recordings, and video without a separate model."
Ember is the fourth model under Meta's Muse brand, after the mid-weight Spark in April, its open-source twin Glimmer in August and the Wonder frontier model in October. Meta released the Muse Agent consumer product in September, and the report described that launch as an unexpected success, with features including Spaces and video chat rolling out earlier in November.