Skip to content
ShilpBot
AI Models

Meta Muse Ember: Pricing, Specs and Limits

Meta Muse Ember: smallest and most efficient Meta model for repeatable agent work, 1M context, 0.10 in / 0.50 out per 1M tokens. Specs, limits and test...

By ShilpBot Team · Published · 3 min read

Muse Ember is Meta's smallest and most efficient model in this series. The listing describes it as designed for complex reasoning execution and the work agents repeat all day: tool calls, coding-agent subtasks, routing and similar steps.

Its cached input price is $0.004 per million tokens, the lowest on this list. If your agent sends the same large prefix again and again, caching can make Ember extremely cheap to run.

Prices and specifications checked against the public OpenRouter catalogue on 2026-11-16. See the live model page for current limits.

Meta Muse Ember quick facts

DetailValue
DeveloperMeta
Model IDmeta/muse-ember
Added to the catalogue2026-11-09
TypeSmallest and most efficient Meta model for repeatable agent work
ParametersNot listed
Context windowabout 1M tokens (1,048,576)
Maximum output943,718 tokens
Input typestext, image, video, file, audio
Price0.10 in / 0.50 out per 1M tokens (cached input 0.004)

Feature support from the catalogue: Tool calling: listed; JSON / structured output: listed; Adjustable reasoning effort: listed.

What Meta Muse Ember is good for

  • Repetitive agent steps such as routing, tool calls and sub-tasks.
  • Workloads with large repeated prompts, thanks to very cheap cached input.
  • Multimodal input across text, images, video, files and audio at low cost.

Limits of Meta Muse Ember

  • Smallest in its family, so expect it to trail Meta's larger Muse models on difficult reasoning.
  • Cache savings only apply if your client reuses the same prompt prefix.
  • Output up to 943,718 tokens is a limit, not a target.

How Meta Muse Ember compares

  • Meta Muse Spark Open: Both have a 1,048,576-token context window; both cap output at 943,718 tokens; it costs 0.10 input and 0.50 output per million tokens against 1.25 and 4.25.
  • Meta Muse Wonder: Both have a 1,048,576-token context window; both cap output at 943,718 tokens; it costs 0.10 input and 0.50 output per million tokens against 4.00 and 20.00.
  • Anthropic Claude Haiku 5.5: It has a larger context window (1,048,576 against 1,000,000 tokens); its output cap is 943,718 tokens against 128,000; it also accepts video and audio input; it costs 0.10 input and 0.50 output per million tokens against 0.10 and 0.50.

Calling Meta Muse Ember from code

Use the model ID meta/muse-ember with any OpenAI-compatible client pointed at OpenRouter:

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "meta/muse-ember", "messages": [{"role": "user", "content": "Explain what you can do in two sentences."}]}'

Prompts that show what Meta Muse Ember can do

  1. Use it as a first-pass classifier on 30 support tickets and compare with a stronger model.
  2. Send the same long system prompt ten times and check whether cached pricing appears on your bill.
  3. Give it a sub-task from a coding agent and measure latency.

What Meta Muse Ember costs

The catalogue lists 0.10 per million input tokens and 0.50 per million output tokens, with cached input at 0.004 per million. Sending 1 million input tokens and receiving 100,000 output tokens would cost about 0.15. A single coding-agent step with a 150,000-token prompt and a 10,000-token answer would cost about $0.020 before caching. If your client reuses the same prompt prefix, cached input at 0.004 per million tokens can cut the input side of that bill by about 96 percent. Prices are set by the provider, can change, and may differ between hosting providers, so check the live model page before you budget.

Keep reading

Compare every paid model by price in our AI model pricing roundup, or see the free models roundup. New to terms like mixture-of-experts? Start with how neural networks learn.

Frequently asked questions

Why is Muse Ember's cached price so low?

The catalogue lists cached input at $0.004 per million tokens against $0.10 for uncached input. Meta prices repeated prompt prefixes very cheaply to favour agent loops.

What are the cost and limits of Meta Muse Ember?

As of 2026-11-16 the catalogue lists 0.10 per million input tokens and 0.50 per million output tokens. About 1M tokens (1,048,576) context window, up to 943,718 output tokens, model ID meta/muse-ember. Prices can change, so check the live model page.

#Meta Muse Ember#Meta model#AI model pricing#AI API pricing#AI chat model#efficient model#OpenRouter