Skip to content
ShilpBot
AI Models

Meta Muse Spark Open: Pricing, Specs and Limits

Meta Muse Spark Open: coding-optimized open-weight model, 1M context, 1.25 in / 4.25 out per 1M tokens. Specs, limits and test prompts.

By ShilpBot Team · Published · 3 min read

Muse Spark Open is a coding-optimized open-weight model from Meta. It is trained specifically for agentic coding workflows, and the listing says it delivers higher first-attempt accuracy, reliable tool calling and deep codebase understanding.

It accepts every input type in the list: text, images, video, files and audio. At $1.25 input and $4.25 output per million tokens it sits above Ember and well below Wonder.

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

Meta Muse Spark Open quick facts

DetailValue
DeveloperMeta
Model IDmeta/muse-spark-open
Added to the catalogue2026-10-15
TypeCoding-optimized open-weight model
ParametersNot listed
Context windowabout 1M tokens (1,048,576)
Maximum output943,718 tokens
Input typestext, image, video, file, audio
Price1.25 in / 4.25 out per 1M tokens (cached input 0.15)

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

What Meta Muse Spark Open is good for

  • Agentic coding on real repositories.
  • Debugging from screenshots, recordings and logs together.
  • Teams that want an open-weight coding model from Meta.

Limits of Meta Muse Spark Open

  • First-attempt accuracy is the vendor's claim. Measure it on your repo.
  • It costs more than coding rivals such as Beam 1.0 or MiniMax M3.1 Flash.
  • Agents can edit files you did not intend, so use version control.

How Meta Muse Spark Open compares

  • Meta Muse Ember: Both have a 1,048,576-token context window; both cap output at 943,718 tokens; it costs 1.25 input and 4.25 output per million tokens against 0.10 and 0.50.
  • Meta Muse Wonder: Both have a 1,048,576-token context window; both cap output at 943,718 tokens; it costs 1.25 input and 4.25 output per million tokens against 4.00 and 20.00.
  • Reflection Beam 1.0: Both have a 1,048,576-token context window; both cap output at 943,718 tokens; it also accepts image and video and file and audio input; it costs 1.25 input and 4.25 output per million tokens against 0.30 and 1.00.

Calling Meta Muse Spark Open from code

Use the model ID meta/muse-spark-open 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-spark-open", "messages": [{"role": "user", "content": "Explain what you can do in two sentences."}]}'

Prompts that show what Meta Muse Spark Open can do

  1. Give it a failing test suite and ask for the smallest passing change.
  2. Ask it to find all call sites of a function in a repo summary.
  3. Run one task on Spark Open and Ember and compare quality and cost.

What Meta Muse Spark Open costs

The catalogue lists 1.25 per million input tokens and 4.25 per million output tokens, with cached input at 0.15 per million. Sending 1 million input tokens and receiving 100,000 output tokens would cost about 1.68. A single coding-agent step with a 150,000-token prompt and a 10,000-token answer would cost about $0.23 before caching. If your client reuses the same prompt prefix, cached input at 0.15 per million tokens can cut the input side of that bill by about 88 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

How does Muse Spark Open compare with Muse Ember?

Spark Open is coding-focused and costs $1.25 input per million tokens. Ember is Meta's smallest, cheapest model at $0.10 input, built for repetitive agent sub-tasks.

What are the cost and limits of Meta Muse Spark Open?

As of 2026-11-16 the catalogue lists 1.25 per million input tokens and 4.25 per million output tokens. About 1M tokens (1,048,576) context window, up to 943,718 output tokens, model ID meta/muse-spark-open. Prices can change, so check the live model page.

#Meta Muse Spark Open#Meta model#AI model pricing#AI API pricing#AI chat model#AI coding model#OpenRouter