Meta’s Muse Ember: A Cheap AI Model That Makes Agent Automation Practical for Small Businesses

What Is Muse Ember?

Meta Superintelligence Lab released a new AI model on Monday called Muse Ember. According to TechCrunch, it is the lab’s lightest and most cost-efficient AI model yet. 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.”

Ember is the fourth model released under Meta’s Muse brand, following the mid-weight Spark (released in April), Spark’s open source twin Glimmer (released in August), and the Wonder frontier model (released in October). The flagship consumer AI product, the Muse Agent, was released in September.

The Price That Changes the Math

At 10 cents per million input tokens, Ember is significantly cheaper than frontier competitors like Anthropic’s Haiku and OpenAI’s Luna, according to TechCrunch. It is also cost-competitive with open source competitors like DeepSeek V4.1 Flash and GLM 5.3 Flash.

For a small business, that price matters. Agent-based automation only pays off when the model running it is cheap enough to work all day. Prices like this put practical automation within reach of businesses that couldn’t justify it a year ago.

How Ember Works in Practice

Given the emphasis on low cost and accuracy at scale, Ember is designed to pair with a larger model like Meta’s Muse Wonder in an agentic harness. As TechCrunch reports, in the case of a customer-support agent, you might see a heavier model establishing rules for how to handle a particular issue, then sending a lighter-weight model to apply the rules across tens of thousands of cases.

As Meta puts it in the announcement, “Muse Wonder plans, Muse Ember handles the volume.” This two-model approach means small businesses can use a powerful, expensive model only for planning and then use Ember for the repetitive, high-volume execution work.

Ember Handles More Than Text

With volume-heavy use cases in mind, Meta also trained Ember to work well across different types of media, including text, images, and audio. According to TechCrunch, the result makes the model surprisingly flexible in what kind of inputs it can handle.

As Meta puts it in the announcement, “agents can read screenshots, charts, and scanned documents, or work from voice memos, meeting recordings, and video without a separate model.” For a small business, this means one model can process customer emails, analyze screenshots of errors, and even listen to voicemails—all without needing separate AI tools.

What This Means for Small Businesses

Ember is built for the work agents repeat all day: tool calls, coding subtasks, routing, extraction, and long document parsing. For a small business, that translates to automating customer support triage, processing invoices, extracting data from contracts, or routing inquiries to the right person.

The cost—10 cents per million input tokens—is low enough that a small business could run thousands of agent tasks per day for a few dollars. Combined with the ability to handle text, images, and audio, Ember makes it feasible to automate workflows that previously required multiple tools or manual effort.

As TechCrunch notes, Meta particularly emphasized Ember’s cost-per-task, using Artificial Analysis’s composite benchmark score to show Ember as cheaper and smarter than competitors. For small businesses watching every dollar, that combination of low cost and solid performance is the key to finally adopting AI agents.

Takeaways

• Muse Ember costs 10 cents per million input tokens, far below comparable models from Anthropic and OpenAI.
• It is designed for high-volume agent tasks like tool calls, coding, routing, extraction, and document parsing.
• Ember can process text, images, and audio, making it a versatile single model for many business automation needs.