Muse Spark 1.2

Meta· released 5 Aug 2026

meta/muse-spark-1.2
Input / 1M tokens
$1.25
Output / 1M tokens
$4.25
Context window1M tokens

Longest answer: 944K tokens

reads textreads imagereads videoreads filetool callingvisionreasoning

About

Muse Spark 1.2 is a Meta reasoning model designed for complex agentic tasks. It accepts text, images, video, and PDF documents, returns text, and has a context window of 1,048,576 tokens. Tools, vision, and reasoning are supported. The maximum output is 943,718 tokens. Input costs $1.25 USD per 1M tokens and output costs $4.25 USD per 1M tokens. Its model ID is meta/muse-spark-1.2; Meta is both the maker and lab. The release date is 2026-08-05. For teams building agentic workflows, the available input types and large context window are key points to assess alongside the token rates. The listed output format is text, even though the input modalities include image and video. The official listing is https://openrouter.ai/meta/muse-spark-1.2.

Who it is for

It suits teams working on complex agentic tasks that need text, image, video, or PDF inputs. Its large context window may be relevant for workflows with substantial context.

What is good

  • Accepts text, images, video, and PDF documents
  • Context window is 1,048,576 tokens
  • Maximum output is 943,718 tokens
  • Tools, vision, and reasoning are supported

What to know first

  • Output is text only
  • Output costs $4.25 USD per 1M tokens

Verdict

Muse Spark 1.2 pairs broad listed inputs with a large context window and reasoning support. Its output is text, and teams should account for the listed input and output rates.

Details

Lab
Metaopenrouter.ai · 3 Oct 2026
Context
1,049Kopenrouter.ai · 3 Oct 2026
Input price
$1.25 / 1Mopenrouter.ai · 3 Oct 2026
Output price
$4.25 / 1Mopenrouter.ai · 3 Oct 2026
Max output
943,718 tokensopenrouter.ai · 3 Oct 2026
Inputs
text, image, video, fileopenrouter.ai · 3 Oct 2026
Open weights
Noopenrouter.ai · 3 Oct 2026

More newest models

See the list