Models·Meta

Muse Spark 1.3: What it does and what it costs

meta/muse-spark-1.3

Muse Spark 1.3 is Meta’s reasoning model for text, images, video, and files.

$1.25Input, per million tokens
$4.25Output, per million tokens
1.05MContext window
944KLongest answer

1.05M tokens of context — about 786,432 words, on a scale that ends at 2M.

What it can do

  • Reasoning
  • Vision
  • Tool use
  • Takes text
  • Takes image
  • Takes video
  • Takes file
  • Gives text

Released 2 Sept 2026 · checked 3 Oct 2026.

About it

Muse Spark 1.3 is a Meta model that can take text, images, video, and files as input and return text. It supports reasoning and tools. Its context window is 1,048,576 tokens, and it can produce up to 943,718 tokens in one answer. Pricing is $1.25 per million input tokens and $4.25 per million output tokens.

Review

In a line

Muse Spark 1.3 is a reasoning model from Meta. It accepts text, images, video, and files, and returns text.

What it's good at

The model supports reasoning and tools, which suit tasks that involve working through information or using tools. It can also take images and video as input, alongside text and files.

Price in plain terms

Input costs $1.25 per million tokens; output costs $4.25 per million tokens. Tokens are pieces of text the model reads or writes. A page of about 500 words costs about 0.081 cents to read. One chat turn with 1,000 tokens in and 500 out costs about 0.34 cents; 1,000 such turns cost about $3.38.

Context and limits

Muse Spark 1.3 can take in up to 1,048,576 tokens at once. That is its context window: the amount of information it can consider in one request. It can return up to 943,718 tokens in one answer. Its output is text.

Ways to use it

The model is available through provider APIs. Its model ID is meta/muse-spark-1.3. Meta is its maker.

Verdict

Consider Muse Spark 1.3 if you need a model that can reason, use tools, and accept more than text, with a large amount of input available in one request. Its listed prices make input cheaper than output, so frequent or lengthy answers may matter more to your bill. If you need the model to return images, video, or files, it is not a fit: it returns text only.

Verdict

Muse Spark 1.3 may suit people comparing models for tasks that involve reasoning, tools, or image and video input. Its large context window can hold substantial information in one request, while its maximum answer is also large. The per-token prices are low in the supplied examples, but output costs more than input. Choose another option if you need anything other than text returned, or if you want a model whose access method is specified beyond provider APIs.

  • Accepts text, images, video, and files as input.
  • Supports reasoning and tools.
  • Context window holds up to 1,048,576 tokens.
  • Can return up to 943,718 tokens in one answer.
  • Returns text only.
  • Output tokens cost more than input tokens.
  • The listed way to access it is through provider APIs.
  • Pick it ifTasks that benefit from reasoning and tool use; Requests involving text, images, video, or files
  • Skip it ifYou need the model to return images, video, or files; You need an access option other than a provider API

Pricing

Input is $1.25 per million tokens and output is $4.25 per million tokens. The supplied examples estimate about 0.081 cents to read a page, about 0.34 cents for one chat turn, and about $3.38 for 1,000 such turns.

Questions

Who makes Muse Spark 1.3?

Meta makes Muse Spark 1.3.

What can Muse Spark 1.3 take as input?

It accepts text, images, video, and files.

What does it return?

It returns text.

How much does it cost?

Input is $1.25 per million tokens and output is $4.25 per million tokens. A typical 1,000-token-in, 500-token-out chat turn is estimated at about 0.34 cents.

How much can it read at once?

Its context window is 1,048,576 tokens.

How can people use it?

It is available through provider APIs. Its model ID is meta/muse-spark-1.3.

On these boards

More from Meta